University of Montana ScholarWorks at University of Montana
Graduate Student Theses, Dissertations, & Professional Papers Graduate School
1999
The effects of body size and morphology on the flight behavior and escape flight performance of birds
Jerred J. Seveyka The University of Montana
Follow this and additional works at: https://scholarworks.umt.edu/etd Let us know how access to this document benefits ou.y
Recommended Citation Seveyka, Jerred J., "The effects of body size and morphology on the flight behavior and escape flight performance of birds" (1999). Graduate Student Theses, Dissertations, & Professional Papers. 6661. https://scholarworks.umt.edu/etd/6661
This Thesis is brought to you for free and open access by the Graduate School at ScholarWorks at University of Montana. It has been accepted for inclusion in Graduate Student Theses, Dissertations, & Professional Papers by an authorized administrator of ScholarWorks at University of Montana. For more information, please contact [email protected]. I
Maureen and Mike MANSFIELD LIBRARY
The University of IVIONTANA
Permission is granted by the author to reproduce this material in its entirety, provided that this material is used for scholarly purposes and is properly cited in published works and reports.
** ** Please check "Yes" or "No" and provide signature
Yes, I grant permission No, I do not grant permission
Author's Signature
Date
Any copying for commercial purposes or financial gain may be undertaken only with the author's explicit consent.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. THE EFFECTS OF BODY SIZE AND MORPHOLOGY ON THE FLIGHT BEHAVIOR AND ESCAPE FLIGHT PERFORMANCE OF BIRDS
by
Jerred J. Seveyka
B A , University of Montana, Missoula, 1996
Presented in partial fulfillment of the requirements
for the degree of
Master of Science
University of Montana
1999
Approved by
Chairr^n, Bo^d of Examiners
Dean, Graduate School
( - X o ~ "7 *7 ______Date
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. UMl Number: EP37462
All rights reserved
INFORMATION TO ALL USERS The quality of this reproduction Is dependent upon the quality of the copy submitted.
In the unlikely event that the author did not send a complete manuscript and there are missing pages, these will be noted. Also, If material had to be removed, a note will Indicate the deletion.
UMT Dissartatton Publishing
UMl EP37462 Published by ProOuest LLC (2013). Copyright In the Dissertation held by the Author. Microform Edition © ProOuest LLC. All rights reserved. This work Is protected against unauthorized copying under Title 17, United States Code
ProQ ^st
ProOuest LLC. 789 East Elsenhower Parkway P.O. Box 1346 Ann Arbor, Ml 4 8 1 0 6 -1 3 4 6
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Seveyka, Jerred Joseph, M.S. Fall 1999 Organismal Biology and Ecology
The effects of body size and morphology on the flight behavior and escape flight performance of birds Director; Kenneth P. Dial V , In general, takeoff or escape flight performance among birds appears to decrease with increasing body size. Two basic groups of models have been proposed to explain the apparent decline in flight performance. The original models suggest that power available for flight decreases due to the effects of wing moment of inertia on wingbeat frequency, while the power required for flight scales nearly independent of mass. However, more recent models suggest that larger birds produce proportionally less lift per unit power output than smaller birds, which results in a decrease in flight performance. I examined the relationship between body size and escape flight performance among wild birds using five species of doves (Family Columbidae) ranging an order of magnitude in body mass (36g to 360g), while accounting for differences in morphology, flight behavior, and phylogeny. Doves were captured in the wild and subsequently video taped as they escaped from a 2.5 m tall, netted tower. From the video and body masses of the birds, I estimated whole- body mass-specific climb power (Pci) during escape flight, this measurement represents the difference between the power available compared to the power required for flight. In addition, I collected morphometric data for 26 species of doves from museum skeleton and wing collections, so that differences in morphology could be compared to documented differences in performance and flight behavior. As mass increased among the doves, wingbeat frequency and mass-specific power output decreased. Wingspan and fight muscle ratio (flight muscle mass divided by body mass) explained nearly all of the variance in escape flight performance. Birds with relatively short wings and high flight muscle ratio, such as the White-tipped Dove, achieved high escape flight wingbeat frequencies and relatively high mass-specific climb power output. A possible tradeoff associated with these attributes is reduced fast-flight performance. Indeed, Mourning Doves, White-winged Doves, and Rock Doves, the most active flyers in the study, had relatively long wings, lower flight muscle ratios and lower escape-flight performance than the other doves studied. The documented positive correlation between flight muscle size and mass-specific climb power is consistent with the predictions from the original and more recent models on the scaling of flight performance. However, although the documented negative correlation between wing length and escape flight performance is consistent the original models on the scaling of flight performance, this result is contrary to the predictions from the more recent models. Furthermore, total flight muscle mass-specific climb power for five species of doves ranged from 95 to 152 Wkg '. This range is surprisingly similar to reported estimates for total power available for flight, despite the fact that my measurements do not include the aerodynamic or inertial components of total power available. Thus, existing estimates of total power available for flight in birds appear to grossly underestimate total power requirements.
II
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. TABLE OF CONTENTS
Page
ABSTRACT...... ii
LIST OF TABLES...... vi
LIST OF FIGURES...... viü
INTRODUCTION...... 1
MATERIALS AND METHODS...... 5
MORPHOMETRICS...... 5
FIELD MORPHOMETRICS...... 5
MORPHOMETRICS OF MUSEUM SPECIMENS...... 6
FIELD OBSERVATIONS...... 7
POWER O U TPU T...... 9
STATISTICAL ANALYSIS...... 12
RESULTS...... 13
MORPHOMETRICS...... 13
FIELD AND ESCAPE FLIGHT OBSERVATIONS...... 15
KINEMATICS...... 16
VELOCITY AND ACCELERATION...... 17
CLIMB POWER OUTPUT...... 18
DISCUSSION...... 20
RELATIONSHIPS BETWEEN MORPHOLOGY AND BEHAVIOR 20
KINEMATICS...... 22
iii
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. FLIGHT PERFORMANCE...... 24
PERFORMANCE VERSUS FLIGHT BEHAVIOR ...... 31
CONCLUSIONS AND FUTURE CONSIDERATIONS...... 34
FUTURE RESEARCH...... 34
REFERENCES...... 36
ACKNOWLEDGMENTS...... 42
TABLES ...... 43
FIGURE LEGENDS ...... 53
APPENDIX A ...... 81
TABLES...... 75
FIGURE LEGENDS...... 81
FIGURES...... 82
APPENDIX B ...... 91
TABLES...... 90
FIGURE LEGENDS...... 91
FIGURES...... 93
APPENDIX C ...... 108
TABLES...... 110
FIGURE LEGENDS...... 111
FIGURES...... 112
APPENDIX D ...... 119
TABLES...... 118
iv
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. FIGURE LEGENDS...... 119
FIGURES...... 121
APPENDIX E ...... 127
FIGURE LEGENDS...... 128
FIGURES...... 129
APPENDIX F ...... 136
TABLES...... 138
FIGURE LEGENDS...... 145
FIGURES...... 147
APPENDIX G ...... 160
FIGURE LEGENDS...... 173
FIGURES...... 174
APPENDIX H ...... ■...... 177
FIGURE LEGEND...... 179
FIGURE...... 180
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. LIST OF TABLES
TABLE Page
Table 1 Morphometric variables for the bones of six species of doves 43
Table 2 Least-squares and Reduced-major-axis regression slopes, and correlation coefficients describing the relationships between log of variable and log body mass for 26 and 6 species of doves 44
Table 3 Morphometric variables for the wings of six species of doves 45
Table 4 Least-squares and reduced-major-axis regression slopes and correlation coefficients describing the relationship between log of flight variable and log body mass for five or six species of doves ...... 46
Table 5 Wingbeat frequencies (Hz) for six species of doves ...... 47
Table 6 Percentage of wingbeat cycle that is downstroke, wingtip reversal or flick phase in six species of doves ...... 48
Table 7 Average velocity and acceleration of doves escaping from the to w er ...... 49
Table 8 Maximum mass-specific, and muscle-mass specific power output for five species of doves ...... 50
Table 9 Regression slopes and correlations coefficients describing the relationships between log of variable and log mass- or muscle mass-specific power output for five species of doves ...... 51
Table 10 Multiple regression models describing the relationships between log variables and log mass-specific climb pow er ...... 52
Appendix A
Table A1 Skeletal measurements of 26 species of doves ...... 75
Table A2 Morphometric variables for the wings of 26 species of doves ......
VI
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Appendix C
Table Cl Regression models describing the relationship between Pectoralis scarring on the keel and flight muscle m ass ...... 110
Table C2 Mean morphological variables used in predicting flight muscle mass from pectoralis scar area ...... 110
Appendix D
Table D1 Mean acceleration magnitudes and angles for the four phases of a wingbeat in three pigeons described separately ...... 118
Table D2 Mean acceleration magnitudes and angles for the four phases of a wingbeat (three pigeons combined) ...... 118
Appendix F
Table FI Morphological variables for the skeleton of nine species of heron .. 138
Table F2 Morphological variables for the wings of nine species of heron ... 140
Table F3 Skeletal measurements from 10 species of kingfisher ...... 141
Table F4 Wing measurements from 6 species of kingfisher ...... 144
Vll
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. LIST OF FIGURES
FIGURE Page
Figure I Method used to measure power output in doves (aka “Tower of Power” ...... 55
Figure 2 Least-squares regression lines and RMA coefficients describing the relationships between selected log-transformed variables and log body mass for six species of doves ...... 56
Figure 3 Least-squares regression lines, RMA and correlation coefficients describing the relationships between log wingbeat frequency and log body mass for six species of doves ...... 60
Figure 4 Least-squares regression lines, RMA and correlation coefficients describing the relationships between selected log transformed variables and log wingbeat frequency for six species of doves 63
Figure 5 Lateral views of takeoff flights illustrating the path of the wingtip and wrist with respect to the body in 5 species of doves ...... 65
Figure 6 Least-squares and RMA regression lines, and correlation coefficients describing the relationship between log mass-specific whole-body climb power and log body mass for five species of doves ...... 68
Figure 7 Least-squares regression lines, correlation coefficients, and RMA regressions describing the relationship between log body mass and log muscle mass-specific power output in five species of doves 69
Figure 8 Least-squares regression lines, RMA and correlation coefficients describing the relationships between selected log transformed variables and log muscle mass-specific power output ...... 72
APPENDIX A
Al Whole-body kinematics of 5 species of doves during takeoff. 82
A2 Proposed phylogeny for the doves in this study ...... 87
Vlll
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. APPENDIX B
E l Scaling of selected morphological variables for 6 species of doves.. 93
B2 Scaling of selected morphological variables for 26 species of doves . 98
B3 Kinematics of doves during fast forward flight ...... 106
B4 Kinematics of doves during slow flight ...... 107
APPENDIX C
C1 Regression model comparing pectoralis scar area to combined pectoralis and supracoracoideus mass for doves ...... 112
C2 Regression model comparing pectoralis scar area to combined pectoralis and supracoracoideus mass for herons ...... 113
APPENDIX D
D 1 Flight path, whole-body kinematics, and acceleration vectors for a pigeon during vertical flight ...... 121
D2 Polar diagrams summarizing acceleration during five wing cycle phases in a pigeon ...... 122
D3 Polar diagrams illustrating changes in acceleration during a wingbeat ...... 123
D4 Polar diagrams illustrating accelerations, body angles, and flight paths during the wingbeat cycle of three pigeons ...... 124
APPENDIX E
El Comparison of lateral kinematics with respect to the body of a Ringed Turtle-dove and Black-billed Magpie during landing and slow flight ...... 129
E2 Comparison of lateral kinematics of a Ringed Turtle-Dove and Black-billed Magpie during landing and slow flight ...... 131
V lll
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. E3 Caudal kinematics of a Ringed Turtle-Dove and Black-billed Magpie during forward flight ...... 133
E4 Lateral kinematics of a Black-billed Magpie during takeoff 134
E5 Kinematics versus wingspan for a Ringed Turtle-Dove ...... 135
APPENDIX F
FI Scaling of skeletal measurements for nine species of heron ...... 147
F2 Scaling of wing measurements for nine species of heron 152
F3 Scaling of skeletal measurements for ten species of kingfisher 154
F4 Scaling of skeletal measurements for six species of kingfisher 159
APPENDIX G
G1 Hypothetical example between sampling rate and kinematics 174
G2 Example of using a grid to scale kinematics ...... 174
G3 Series of tracing of a human walking ...... 175
G4 Sequential images of a Ringed Turtle-Dove during flight ...... 175
G5 Example of presenting movement with respect to the body 176
G6 Example of presenting movement over time ...... 176
APPENDIX H
HI Scaling of level flight wingbeat frequency in herons ...... 180
IX
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. INTRODUCTION
In general, takeoff and burst performance among flying birds and other flying
animals appear to decrease with increasing body size. Typically, small birds are highly
maneuverable, can hover and can readily vary flight speeds, while large birds are less
capable of these feats. The explanations for this disparity in flight performance have
generally been associated with an adverse scaling of power available versus the power
required for flight; that is, as body mass increases, the excess of power available for flight,
which could be used to increase velocity, takeoff angle, or acceleration decreases
(Pennycuick 1968, 1969, 1989; Lighthill 1977; Weis-Fogh 1977). However, experiments
in which flying birds carried a payload, non-sustained mass-specific (i.e., per kilogram of
body mass) lift production and non-sustained, mass-specific induced power have been
estimated to scale nearly independent of size (Marden 1987, 1990, 1994; Ellington 1991)
The mass-specific power required for animal flight has been predicted to scale
between mass (M)” and (Ellington 1991). This prediction is based on the scaling of
induced power (the power required to generate lift and thrust), and assumes that induced
power dominates the power required for slow flight (Pennycuick 1968; Marden 1987;
Ellington 1991). Mass-specific induced power is proportional to the square root of wing
disk loading (weight divided by the area of the disk swept by the wings); thus, animals
with relatively longer wings have relatively lower induced power requirements
^ (Pennycuick 1968, 1969, 1975; Epting and Casey 1973; Ellington 1984). Among
geometrically similar flying organisms, wing loading (weight divided by total wing area)
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 2
and wing disk loading increase as thus, according to this model, the mass-specific
power required for flight should scale as M*''®. However, such models may not adequately
predict the scaling of the power required for burst flight because they neglect the scaling
of inertial power (the power required to oscillate the wings) and profile power (the power
required to overcome the pressure and friction drag of the wings), which are likely to be
high during slow flight when wingbeat fi-equencies and amplitudes are high (Weis-Fogh
1972; Norberg 1990; van den Berg and Rayner 1995).
While the mass-specific power required for flight is predicted to be independent of
mass or increase slightly as mass increases, the mass-specific power available for flight is
predicted to decrease due to the negative scaling of wingbeat frequency (Pennycuick
1969, 1972). Hill (1950) proposed that the scaling of limb inertia and the properties of
connective tissue limit the scaling of the maximum attainable frequency for an oscillating
limb, such that maximum limb oscillation frequency scales as among geometrically
similar organisms. This scaling coefficient has been reported across a broad size range of
distantly related birds (Van Den Berg and Rayner 1995), among woodpeckers (Tobalske
1996), and among Galliformes during burst flight (Tobalske and Dial 1996b). If muscle
stress (the force divided by the cross-sectional area of the muscle), and strain (the change
in muscle length divided by the initial muscle length), are independent of mass, then the
mean power output per unit mass of muscle should be proportional to the frequency of
muscle contraction (Hill 1950, Pennycuick 1969, 1972). Furthermore, if the flight muscle
mass represents a constant fraction of the total body mass, wing inertia scales as predicted
by geometric similarity, and the flight muscle fiber composition and the duration of force
production during a wingbeat cycle are independent of body mass, then the maximum
mass-specific power available for flight should be proportional to the maximum wingbeat
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. frequency (Pennycuick 1975). Birds with relatively shorter wings, larger flight muscles,
and relatively longer downstroke durations are predicted to have higher mass-specific
power available due to their relatively higher maximum wingbeat frequencies, force
production, and duration of force production, respectively. Therefore, if the mass-specific
power required for flight scales independent of mass, then the birds with relatively larger
flight muscles and higher wingbeat frequencies should have the highest takeoff
performance.
Although mass-specific power output among flying animals is predicted to
decrease due to the negative scaling of wingbeat frequency, the empirical data on the
scaling of power output among flying animals vary considerably. During escape flight,
takeoff acceleration was found to scale negatively among small passerine birds (Dejong
1983) and aerial insectivores (Warrick 1998). These data suggest that maximum mass-
specific power output scales negatively and possibly as predicted by the arguments of
geometric similarity. Similarly, maximum mass-specific takeoff power among four species
of galliforms scales as (Tobalske and Dial 1996b).
However, two other studies have documented positive scaling coefficients for
mass-specific power output in animals flying with added mass. Among three species of
trained doves carrying added mass vertically, mass-specific climb power scales as M”
(Bosdyk and Tobalske 1999). In addition, Marden (1987) determined that maximum
lifting force production was proportional to flight muscle mass across a broad range of
insects, birds and bats based on the maximum weight that an animal could carry and still
achieve flight. Using these data and data for the load carrying capacity of Harris’ Hawks
(Parabuteo unicinctiis) reported by Pennycuick et.al. (1989), Marden (1990) indirectly
estimated the maximum induced power output for the flying animals using the actuator-
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. disk equation (Pennycuick 1968; Weis-Fogh 1972; Alexander 1983). From this analysis,
maximum body mass-specific induced power output is reported to scale as M" among
flying animals (Marden 1990). Ellington (1991) reexamined these data and reported that
muscle mass-specific power scaled as across all of the combined flying animals and
as among 10 species of birds (12 individual birds were measured); however, these
exponents were not significantly different from zero. Based on Marden’s (1987) maximum
lift force data, the noticeable decrease in flight performance with increasing mass is
believed to be the result of the adverse scaling of lift production per unit power output
(Ellington 1991; Marden 1994); yet this interpretation is based on indirect estimates of
induced power output across a wide range of distantly related animals. Among
hummingbirds maximum mass-specific lifting ability was found to increase with increasing
body mass (Chai and Millard 1997).
Many of the investigations of the scaling of flight behavior or morphology have
been conducted at a broad taxonomic range, but phylogenetic relationships were not
controlled or accounted for (e.g., Greenwalt 1962, 1975; DeJong 1983; Scholey 1983;
Pennycuick 1990, 1996; Marden 1987, 1990, 1994; Van Den Berg and Rayner 1995).
These models and data may be too uncontrolled to adequately explain performance across
smaller size ranges, particularly within a taxon. Although a few important variables may
explain most of the variation in performance at very broad scales, numerous variables may
be important for determining the power required or power available for escape flight.
What has been lacking in the determination of the scaling of power output among flying
birds is a phylogenetically controlled measure of climb power that considers the numerous
variables that might influence flight performance.
Although investigation across broad taxonomic levels and size ranges (e.g., small
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. insects [.019g] to large bats and birds [920g] [Marden 1987, 1990, 1994; Ellington 1991])
may provide information on the ultimate physical constraints that determine the scaling of
flight performance, they may overlook the important biological factors influencing flight
performance, such as selection, adaptation, or phylogenetic inertia. Several recent
investigations of flight behavior and performance in birds have controlled for the variation
in morphology, ecology, and phylogenetic history (e.g., Tobalske 1996; Tobalske and
Dial 1996b; Warrick 1998; Bosdyk and Tobalske 1999).
The primary purpose of this study was to examine the relationships between
escape flight performance and morphology, flight behavior, and kinematics across a size
range of closely-related, wild birds from the family Columbidae (doves and pigeons). To
document escape flight performance, climb power was measured from video of doves
escaping vertically from an enclosure. Assuming that mass-specific power required for
flight scaled independent of mass for the birds, then maximum mass-specific climb power
should represent the difference between power available for flight and the power required
(i.e., climb power is the excess of power available beyond that required). The models
produced from this study were then compared to the existing models of flight
performance. A secondary goal of this study was to examine possible relationships
between the ecology of selected doves and their morphology, flight behavior, and flight
performance.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. METHODS
Morphometries
Field Morphometries
Body mass and four morphological measurements were collected from birds
captured in the field. Rock Doves {Cohimba livid) were captured using an electronically
remote-controlled trap in Missoula, Montana and all other doves were captured using mist
nets in the Rio Grande Valley of southern Texas. The Rock Doves were weighed to ±
3.0g on a spring scale and all other doves were weighed to ±1 .Og on a 50-g or 300-g
Pesola spring scale. Handling time of birds was minimized by using a video-recorder to
document spread wing dimensions. One wing (stretched as in mid-downstroke of flapping
flight ) from each bird was videotaped (Hi-8 format, Sony Model 910) outstretched across
a clipboard decorated with lines of known length (15-30cm). Video images of each bird
were downloaded onto a computer (Macintosh Quadra 950 using Screenplay, Apple, Inc.)
and measurements were obtained using Image 1.6 (National Institutes of Health).
Morphological measurements obtained were: single-wing area (a), single-wing length (/),
wing-root chord (c), and body width (6). I calculated total wing area (S) as 2a+ch,
wingspan (L) as 2l+b, wing loading as total wing area divided by weight (w), wing disk
loading as w*4*tt’^*L'^), and aspect ratio as L^*S'\
Morphometries of Museum Speeimens
Seventeen morphological measurements were obtained from museum specimens
for 26 species of Columbids (Tables 1 and Al). The lengths of the sternum, fiircula,
coracoideus, scapula, humerus, radius, ulna, and manus were measured as in Tobalske
(1996). In addition, pectoralis size was estimated by tracing the area of origin of the
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. pectoralis from scarring on the keel; this estimate provides an index of muscle size only
and does not represent the physiological cross-sectional area of the bipennate pectoralis
muscle (Alexander 1983). The traced areas were then digitally scanned onto a computer
and measured using NIH Image 1.6. External wing measurements were acquired from
museum specimens of dried, spread wings in a similar manner as described previously for
wild caught birds. Body widths were estimated from the least-squares regression line fit
to the log of body width against the log of body mass from doves captured in the field.
The body width estimates were multiplied by the measured base wing chord to estimate
the area between the wings, which was used in calculating total wing area, aspect ratio,
wing loading, and wing disk loading [due to the potential error in extrapolated estimates
of body width, an aspect-ratio index was also calculated sls P * a \ similar to Tobalske
(1996)]. Body masses were either recorded from museum tags associated with the dried
skeletons and spread wings, or, when these data were not available, an average body mass
reported for the species from Dunning (1993) was used.
Field Observations
Rock Dove flights used for analysis were recorded in Missoula, Montana
(elevation 975m); flights of other dove species were recorded in the Rio Grande Valley of
southern Texas (elevation less than 50m). Flight kinematics were collected using a high
speed video camera (60 fields s ', Hi-8 format, Sony Model 910). The Hi-8 video was
transferred to S-YHS and a time code (Horita II model TG 50) was added. Video was
viewed and analyzed using a Panasonic AGI 960 editing video player. Flights of doves
during level, non-maneuvering, uninterrupted bouts of flapping flight in calm air (wind <
1ms ') and during takeoff (the first five full wingbeats of ascending flight) were included
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 8
in the analysis of wingbeat frequency, which was determined by dividing the number of
wingbeats recorded in a flight by the elapsed time (determined using the number of frames
elapsed).
Wing kinematics were recorded from birds that were flying toward or away from
the camera (henceforth cranial or caudal view) or lateral to the camera’s field of view
(henceforth lateral view). Takeoff and ascending flight kinematics of Common Ground
Doves {Columbinapasserind), Inca Doves {Columbina inca). Mourning Doves {Zenaida
macrourd). White-winged Doves {Zenaida asiaticd), and White-tipped Doves (Leptotila
verreauxi) were recorded in Texas at 250 Frames s'' or 500 Frames s"' using a high-speed
digital video camera (Redlakes Motionscope 2000). Ascending flight kinematics of
trained Rock Doves were recorded at 500 Frames s'* (Redlakes Motionscope 2000). The
kinematics presented describe the path of the wrist and wingtip with respect to the body of
the flying bird. Movements were traced directly from a video monitor, then digitally
scanned onto a computer (Quadra 950) and retraced using a graphics software program
(Canvas 3.5, Deneba Software Inc.).
High speed video of doves during takeoff (viewed from lateral, cranial or caudal
views) provided estimates of the duration of three phases of the wingstroke cycle and the
relative portion of each phase in a complete wingbeat cycle . The three phases of the
wingbeat cycle considered were: downstroke, wing turn around, and the flick phase,
which were chosen due to the possibility that each phase provides separate useful forces
during the wingbeat cycle (Brown 1948, 1963; Scholey 1983; Seveyka and Warrick, in
prep). Downstroke was defined as the portion of the wingbeat cycle from wing extension
at the top of upstroke until the beginning of wrist retraction at the bottom of the wingbeat
cycle. From this point the wing was considered to be in wing turn around until the
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. wingtip and wrist returned to the same elevation as each other during upstroke. The flick
phase was defined as the period between the end of wing turn around and the beginning of
the next downstroke. Measurement error for determining the timing of full wingbeat cycle
kinematic events is determined by wingbeat frequency and frame rate, and is likely lowest
for birds with low wingbeat frequencies (Scholey 1983). Confidence in the measurements
was, therefore, higher for the larger doves than for the smaller doves.
Whole-Body Climb Power Output
Six species of Columbids (Common Ground Dove, Inca Dove, Mourning Dove,
White-winged Dove, White-tipped Dove, and Rock Dove) were captured in the wild.
Following morphometric measurements, birds were allowed to escape from a 2.5m tall,
Im^ frame covered with netting on all sides, but with an open top (hereafter referred to as
the "tower"; Figure I). The escape flights were recorded on video tape using two
cameras. One camera (approximately 8m from the mid-section of the tower, Panasonic
S-VHS; 60 fields s*^) recorded the entire flight of the bird escaping, and the second camera
(3m from the tower and perpendicular to the first camera, Sony Model 910; 60 fields s'*)
recorded the wing kinematics and position of the birds as they flew through the tower.
Because the Common Ground Doves either escaped through the mesh of the tower or
failed to fly vertically out of the tower, I was unable to determine the climb power output
for this species.
Trials using Rock Doves were conducted in Missoula (temperatures of
approximately 17° C) and all other trials were conducted in Southern Texas (temperatures
of approximately 32° C). Based on the temperature and elevation of the two sites, average
air density was estimated to be 5% higher at the Texas sites. This difference in air density
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 10
changes the estimated induced power (based on the momentum jet theory) by less than
5%. Because the difference in air density between the two sites was slight and the
estimated difference in induced power was small, I elected not to correct for the effects of
air density.
From the flight video of the birds escaping, I determined each bird’s average
acceleration to peak velocity, as well as the distance covered and the time elapsed for the
bird to move from the second wingbeat of the flight until the bird reached peak velocity.
The velocity, acceleration, distance and time estimates were used to calculate maximum,
whole-body climb power output (Pci) of each bird using the following formula for power
(note: the acceleration values used in the formula were the addition of the acceleration due
to gravity (9.81 ms'^) and measured average acceleration to peak velocity):
Whole-body climb power output = m * fg + al * L t
(m = body mass; g = gravitational acceleration =9.81 ms-2; a = whole-body acceleration
to peak velocity; L = distance moved; t = time elapsed). Whole-body climb power output
as measured here is not equivalent to total climb power, which is the combination of the
whole-body component measured here and the aerodynamic power requirements for
vertical flight (induced, profile, parasite and inertial power requirements; Ptot = Paero +
Pci). However, I will refer to the whole-body climb power measurements produced by the
above formula simply as climb power (Pci). I divided climb power by the mass of the bird
to calculate maximum mass-specific climb power (Wkg"'), or by the estimated muscle
mass of the bird from Hartman (1961) and from pectoralis scarring on the keel to
determine muscle mass-specific climb power output (Wkg‘‘). My discussion of muscle
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 11
mass-specific climb power will be restricted to combined pectoralis and supracoracoideus
mass-specific climb power. I chose this measurement because it probably provides the
most realistic estimates of muscle mass-specific climb power output during escape flight.
Pectoralis mass-specific climb power probably over-estimates the muscle mass-specific
climb power during burst flight due to activity of other flight muscles during takeoff flight
(Dial 1992a, 1992b), and, conversely, the all up flight muscle mass-specific climb power
(climb power divided by the total flight muscle mass; pectoralis, supracoracoideus, and
intrinsic wing muscles) probably underestimates the muscle mass-specific climb power
output because not all of the muscles of the wing are likely to be recruited equally during
escape flight.
To measure climb power, video of escape flights were projected (liquid crystal
video projector) onto a white dry eraser board and the position of a bird’s head in the
tower was plotted every 6 frames (0. Is). A preceding or succeeding frame was used if the
head was more visible in that frame. Distances between points were measured to the
nearest millimeter using a ruler, and converted to actual distance by scaling the measured
flight distances to the measured height of the tower (known to be 2.5m in height). For
most flights the measured height of the projected tower was between 0.7m and 0.97m. If
the bird flew along the front of the tower, the height of the front of the tower on the video
image was used as the scale. In most cases the birds flew in the middle portion of the
tower, so the scale was based on the average between the measured height of the back of
the tower (furthest from the camera) and the front of the tower. Measurement error for
acceleration was estimated by re-measuring several flights on separate occasions (3-5
times), and determined to be approximately 12%, but measurement error for mass-specific
power was approximately 3%. In addition to possible errors of measurement, error may
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 12
have arisen due to the bird moving toward or away from the camera while in the tower;
however, the birds were generally released so that they were in the middle of the tower
and flying lateral to the camera. When released, most birds flew straight up, but a few
birds may have flown 0.5m toward or away from the camera during their escape.
Statistical Analysis
To reduce the risk of pseudo replication from measuring a single bird more than
once, different geographic locations (areas separated by more than 3 kilometers) were
used as sample units for field data, unless several doves were recorded simultaneously, in
which case, each bird was considered a sampling unit. However, White-tipped Doves
were all netted in the same area over several weeks time, so each dove, after the first
dove, was marked after being caught (the first White-tipped Dove caught was not marked
and there is a chance that it was recaptured; however, based on the morphometries, it does
not appear that the same dove was measured twice). For morphological measurements
individual birds were considered a sampling unit. It is important to note that the
relatedness between individual birds of a species measured is unknown and is assumed to
be low; if the birds were closely related, then the power of any statistical test may be
lowered depending on the extent of the relatedness.
All variables were log-transformed and analyzed using least-squares linear
regression models to examine the relationships between body mass, morphology, power,
and flight behavior (SPSS 6.0). In addition, correlations coefficients and reduced-major-
axis (RMA) regression coefficients were calculated to account for the presence of
measurement error and statistical variation in both the independent and the dependent
variables (Sokal and Rohlf 1981; Rayner 1985). RMA regression coefficients were
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 13
calculated by dividing the least-squares linear regression coefficient (computed using
Cricket Graph III 1.5, Comp. Assoc. Int., Inc., 1992), by the correlation coefficient (r).
Takeoff and escape flight wingbeat frequencies within a species were compared using a
Student’s t-test (Excel 4.0, Microsoft, Inc.). Maximum mass-specific power output was
also compared between species using a t-test assuming unequal variance within the
populations (Excel 4.0, Microsoft, Inc.),
I used the independent-contrasts method (Felsenstein 1985; Garland et. al. 1992)
to correct for statistical non-independence of the species examined using PDTree software
(Jones and Garland 1993). Branch lengths for the phylogenetic relationships among the
species in this study are from Sibley and Alquist (1990; Appendix A, Figure A2). In
general, the scaling relationships did not change after correcting for relatedness, therefore,
corrected values are presented only when the coefficients changed considerably.
RESULTS
Morphometries
Among 26 species of doves and among the 6 species in this study, pectoralis scar
area (Figure 2A), and the lengths of the ulna (Figure 2B), sternum, radius, humerus,
manus (Figure 2C), coracoideus, scapula, and furcula all scaled nearly isometrically
(Tables 1 and 2, Appendix B, Figures B1 and B2), in which log-transformed square
dimensions scale proportional to M”^’ and log transformed linear dimensions scale
proportional to IVf (note that the White-tipped Doves studied in the field were from
northern-most populations and on average 13% larger in mass than the museum
specimens). Correlation coefficients for these scaling relationships ranged from 0.92 to
0.99 for the 26 species comparisons, and from 0.98 to 0.99 for the 6 species comparisons.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 14
indicating a strong linear association in the log-transformed data (Table 2).
Pectoralis scar area was measured as an index of flight muscle mass. Log
pectoralis scar areas for 8 dove species were highly correlated with log transformed
pectoralis mass (r^ = 0.98, p < 0.001), combined pectoralis and supracoracoideus mass (r^
= 0.978, p < 0.001), and total flight muscle mass (r^ = 0.974, p < 0.001) for the same
species from Hartman (1961). Pectoralis, combined pectoralis and supracoracoideus, and
total flight muscle ratios used in this study for the Rock Dove (20%, 23.5% and 31%,
respectively) and White-tipped Dove (24.7%, 30%, and 36.9%, respectively) were taken
directly from Hartman (1961). The same series of flight muscle ratios used for the Inca
Dove (22.7%, 28%, and 34%, respectively) were averages from similar sized doves of the
same genus also reported in Hartman (1961); these values correspond well to the output
of the pectoralis scar area to flight muscle mass models described above. Finally, the same
series of flight muscle mass ratios used for the Mourning Dove and White-tipped Dove
were assumed to be 23.5%, 29%, and 35%, respectively; these values also correspond
well with the above model and the measured pectoralis scar area.
Total wing area and wing length scaled slightly more positively than predicted by
geometric similarity for both the 26 species comparisons (RMA slopes of 0.75 and 0.40,
respectively; Table 2, Appendix B, Figure B2) and the 6 species comparisons (RMA
slopes of 0.73 and 0.41, respectively; Table 2, Figure 2D). Wingspan and wing length
among the 6 species also scaled more positively than predicted by geometric similarity and
more positively than the scaling of the lengths of the bones of the wing (Table 2, Figure
2E), indicating that the lengths of the primaries increase disproportionately with mass.
Aspect-ratio was not highly correlated with body mass for the 6 species comparison
(RMA = 0.08; r = 0.58, p = 0.20; Figure 2F), but it scaled slightly positively with a
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 15
significant correlation coefficient for the 26 species comparison (RMA = 0.14; r = 0.51, p
< 0,05; Table 3, Appendix B, Figure B2). The Rock Dove had the highest aspect ratio
(6.6) among the 26 species of doves measured, but the aspect ratios of the White-winged
Dove (5.7) and Mourning Dove (6.1) were also high compared to the overall average of
the 26 dove species (5.3; Table 3, Appendix A, Table A2, Appendix B, Figure B2). Low
aspect ratios were documented in the Inca Dove, Common Ground Dove and White-
tipped Dove (5.0, 5.3, and 4.9, respectively; Table 3).
Wing loading scaled nearly isometrically among the 26 species of doves (RMA =
0.34), but scaled less positively among the 6 species comparison (RMA = 0.27; Figure 20,
Table 2). Wing disk loading scaled positively for the 26 species comparison (Table 2,
Appendix B, Figure B2), but was independent of mass in the 6 species comparison (Table
2, Figure 2H). Due to its short wingspan, wing disk loading of the White-tipped dove
(12.2 Nm"^) was very high compared to the other doves studied (Table 3); among the
doves, wing disk loading increased with increasing mass (RMA = 0.265; r = 0.996), but
not as rapidly as expected for geometrically similar birds (M^''^).
Field and Escape Flight Observations
Wingbeat frequency during takeoff and escape flights decreased with increasing
body mass as predicted by geometric similarity (RMA slopes of -0.29 and -0.35,
respectively; Table 4, Figure 3 A and 3B). Average takeoff wingbeat frequency was
highest in the Common Ground Dove (17.5 Hz), and the Inca Dove (16.4 Hz) and lowest
in the Rock Dove (8.9 Hz) (Table 5). Escape flight wingbeat frequencies of doves
released into the tower were significantly higher than takeoff wingbeat frequencies
observed in the field for the same species, indicating that all of the species were highly
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 16
motivated to escape (Table 5),
Characteristics, such as wing length, total wing area, wingspan, and lengths of the
ulna, radius, humerus, and manus were all negatively associated with wingbeat frequency
(r^ from 0.98 to >0.99, p < 0.005; Table 6, Figure 4). Take-off and escape flight
wingbeat frequencies were expected to decrease directly with wing length, wingspan, or
skeletal element length (Hill 1950), but in all cases wingbeat frequency decreased less
rapidly than expected for these variables (Table 4, Figure 4). Level flight wingbeat
frequencies among the 6 species examined in this study were lower than during takeoff
flight and scaled as (Figure 3C, Table 4) rather than predicted by Pennycuick
(1975). Scholey (1983) reported wingbeat frequencies of 7.54 Hz for 202g Collared
Doves {Streptopeha decaocto), and 5.71 Hz for 447g Wood Pigeons {Colwnba
palumbus), which correspond well with the values reported here (Figure 4C, Table 5).
Wingbeat frequencies not only varied between modes of flight, but also varied during
individual bouts of level flight within an individual bird. For example, during level flight
bouts in Inca Doves and Common Ground Doves wingbeat frequency varied between 8
and 16 Hz, or 11 and 18 Hz, respectively.
Kinematics
During takeoff flights all of the dove species exhibited a wingtip reversal upstroke
and awing clap (Weis-Fogh 1973; Lighthill 1973; Scholey 1983) at the
upstroke/downstroke transition (Appendix A, Figure Al). In a wingtip reversal upstroke
the distal wing is strongly supinated during the upstroke and the wingtips and wrist are
held away from the body. From a lateral view the path of the wingtip and wrist during
upstroke with respect to the body are brought craniad to the downstroke path (Brown
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 17
1948, 1963; Scholey 1983; Tobalske and Dial 1996a, 1996b). This type ofwingstroke
cycle may not only produce lift during the downstroke, but also yield useful force during
the upstroke (Brown 1948, 1963; Norberg 1990; Aldridge 1986; Seveyka and Warrick, in
prep). Furthermore, the clap-and-fling wingstroke exhibited by the doves may increase lift
production through non-steady aerodynamic mechanisms (Weis-Fogh 1973; Lighthill
1973; Ellington 1984).
There was no evidence to suggest that any of the dove species were extending the
duration of any phase of the wingstroke cycle in order to change the duration of force
production. Relative durations of each phase of the wingstroke cycle (downstroke,
wingtip reversal, and flick phases) scaled independent of mass. On average the
downstroke, wingtip reversal, and flick phases occupied approximately 52%, 26%, and
22% of the wingstroke cycle, respectively (Table 6).
An additional kinematic event that was documented in all of the species studied
was a rapid rolling and wing-clapping behavior, wherein the doves rolled to one side
during flight and clapped both wings together. Although this behavior was recorded more
frequently when doves were flushed (most notably in Common Ground, Inca, and
Mourning Doves), it was also witnessed in birds that were flying straight and level (most
often in White-winged, Mourning and Rock Doves).
Velocity and Acceleration
In general, average escape velocity and acceleration to peak velocity decreased
with increasing mass (Table 7). Average velocity (mean ± SE) during escape flight was
highest in the Inca Dove (3.4 ± 0.2 ms''), and lowest in the Rock Dove (2.4 ± 0.2 ms'*;
Table 7). Average acceleration to peak velocity was highest in the White-tipped Dove
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 18
(4.8 ms‘^; Table 7).
Climb Power
Average mass-specific climb power output during escape flight was highest in the
Inca Dove (50.5 ±3.0 Wkg'^; mean ± S) and White-tipped Dove (47.1 ± 5.0 Wkg’^), and
lowest in the Rock Dove (29.0 ±1.1 Wkg'*; Table 8). Maximum mass-specific climb
power output scaled negatively (Least squares regression = -0.24, P ~ 0.69, p = 0.081,
RMA = -0.29; Figure 7, Table 9). Phylogenetic correction for this comparison yielded a
RMA regression slope of -0.31 (Least squares regression = -0.248, r = 0.857, p = 0.143;
Figure 6). Muscle mass-specific climb power output declined less rapidly, but more
significantly with increasing body mass than mass-specific climb power output (Figure 10,
Table 9). If the pectoralis was entirely responsible for force production during flight, then,
based on climb power output measurements, average pectoralis mass-specific climb power
output would range from 228 Wkg‘‘ in the Inca Dove to 148 Wkg‘‘ in the Rock Dove
(Table 8). The highest values approach the predicted anaerobic maximum for vertebrate
muscle (Weis-Fogh and Alexander 1977; 250 Wkg'*), and maximum instantaneous power
output that has been measured for muscle from a variety of animals (200 -300 Wkg'\
Franklin and Johnson 1997). However, including the supracoracoideus mass in the
estimates of muscle mass-specific climb power yield values ranging from 185 Wkg'' in the
Inca Dove to 126 Wkg'' in the Rock Dove (Table 8). Marden (1990, 1994) and Ellington
(1991) considered total flight muscle mass (pectoralis, supracoracoideus, and intrinsic
wing musculature) in their calculations of muscle mass-specific induced power. Following
similar logic, total flight muscle mass-specific climb power ranged from 95 Wkg'' in the
Rock Dove to 152 Wkg ' for the Inca Dove (Table 8).
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 19
All of the models indicate that mass-specific climb power is positively associated
with flight muscle size and negatively associated with wing length. Neither mass-specific,
nor muscle mass-specific climb power were correlated with wing loading or wing disk
loading (Table 9); however, log mass-specific climb power output was strongly correlated
with the log of flight muscle size, wingspan, wing length, and wingbeat frequency (Table
9). Of the measured factors, log mass-specific climb power was most strongly correlated
with log wingspan and the log of combined pectoralis and supracoracoideus flight muscle
ratio (Table 9). A multiple regression model using these factors explained all of the
variance in the data (r^ = 1.0, p ^ 0.001; log mass-specific climb power = 0.44 +1.22 (log
combined pectoralis and supracoracoideus flight muscle ratio) - 0.244 (log wingspan)-.
Table 10). This model includes assumptions on the flight muscle ratios of Mourning
Doves and White-winged Doves, nevertheless changing the pectoralis and
supracoracoideus flight muscle ratio for these species to 30% or 27% did not seriously
affect the model (Table 10). In addition, substituting wing length for wingspan did
substantially affect the model (Table 10). An alternative model using pectoralis scarring
area instead of flight muscle ratio yields similar relationships between the variables, but
different coefficients {log mass-specific climb power = 3.91 - 1.51 (log wingspan) + 0.59
(logpectoralis scar area)-, r^ = 0.975, p = 0.025; Table 10). Log muscle mass-specific
climb power was negatively correlated with the log of wing length, wingspan, and lengths
of the ulna, radius, humerus, and manus (Table 9). All of these factors were strongly
correlated with log wingbeat frequency, which was also strongly correlated with the log of
muscle mass-specific climb power (r^ = 0.972, p = 0.002).
DISCUSSION
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 20
Among the six species of doves in this study, morphology, wingbeat frequency and
climb power output were strongly associated with body mass, but there were no observed
associations between body mass and measured wingbeat kinematics. Wingbeat frequency
among the doves scaled with mass as predicted despite being less negatively
associated with wing length, wingspan and wing bone lengths than predicted (predicted M‘
\ Hill 1950; Pennycuick 1975). As expected, mass-specific climb power among five
species of doves was positively associated with flight muscle ratio and negatively
associated with wing length. Most notably, climb power output for the doves spanned a
similar range as previously reported values for total power available for flight in birds (95
to 152 Wkg'^ reported here versus 111 to 177 Wkg'^ reported by Ellington [1991] and
Marden [1994]), even though climb power (Pci) probably represents only a fraction of the
total power available for flight. This observation suggests that the most recent estimates
of power available for flight in birds (Ellington 1991; Marden 1994) grossly
underestimates total power available.
Relationships between morphology and flight behavior
The wing and skeletal dimensions documented for the doves studied scaled nearly
isometrically, and similar to previously reported scaling trends among broad surveys of
flying animals (Greenwalt 1962; Rayner 1988; Norberg 1990; Table 2, Figures 2).
However, the slopes of the lines describing the relationship between log body mass and
the log of wing length, wingspan, and the lengths of the bones of the wing were all slightly
higher than expected by geometric similarity (Table 2, Figures 2). Similar deviations have
been observed in other groups of closely related birds (Norberg 1990; Tobalske 1996).
Wingspan and wing length were relatively longer in species that were active flyers and/or
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 21
migratory (Rock Dove, Mourning Dove, and White-winged Dove), than in more
sedentary, non-migratory dove species (Inca Dove and White-tipped Dove; Figure 2,
Table 2). In addition, the active flyers had higher aspect ratios than the less active species
(Figure 2, Table 2). The longer, high aspect ratio wings of the Mourning, White-winged
and Rock Doves probably increase efficiency and reduce flight costs during long or
frequent flights (Pennycuick 1989; Norberg 1990).
Takeoff, level flight, and escape flight wingbeat frequencies among six species of
doves scaled similar to predicted values for geometrically similar animals (M‘*^^; Hill
1950). Similar patterns have been reported across a wide range of flying animals
(Greenwalt 1962; Norberg 1990), in woodpeckers (Tobalske 1996), and in the
Galliformes (Tobalske and Dial 1996b). However, this relationship was not expected
among the doves because wingspan, wing length, and wing bone lengths increased
disproportionately with mass (Figure 2, Table 2). Based on the scaling of wing
morphology, wingbeat frequency was predicted to decline more rapidly with mass than
expected for geometrically similar animals (i.e., < L'*; Hill 1950). Since wingbeat
frequency did not decline as M '\ and muscle mass did not increase disproportionately with
wing length, the most plausible explanation for the scaling of wingbeat frequency among
the doves is that wing moment of inertia does not increase proportionally with wing length
(Hill 1950; Pennycuick 1968; van den Berg and Rayner 1996). Among the doves, wing
length increased faster than the lengths of the wing bones, thus the disproportionate
increase in wing length with mass was due to a disproportionate increase in the lengths of
the primaries. Since feathers are light relative to bone, an increase in wing length may not
yield predicted increases in the moment of inertia of the wing. Thus, birds with relatively
long wings may benefit from having a long wing (e.g., lower induced power requirements;
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 22
Pennycuick 1968), without dramatically increasing their inertial power requirements.
Indeed, a pigeon wing stripped of its feathers maintains 65% of its moment of inertia (van
den Berg and Rayner 1995). If an increase in wing length due to an increase in primary
length does not yield a proportional decrease in wingbeat frequency, then a bird may
benefit from maintaining relatively high maximum wingbeat frequency while lowering
induced power requirements (Pennycuick 1968; Rayner 1979; Norberg 1990). However, a
long wing should have high extrinsic inertia (inertia of the air mass accelerated by the wing
during flapping). Extrinsic inertia probably restricts wingbeat frequencies in coursing
insectivores (Warrick 1998), but its role in determining wingbeat frequencies among the
doves is unclear. Future studies on maximum wingbeat frequency and wing moment of
inertia within a closely related group of birds may provide further insight into the subtle
relationships between wing morphology and maximum wingbeat frequency.
Kinematics
Although variation in the wingstroke cycle could result in variation in measured
performance (Marden 1987, 1994; Ellington 1991), all of the dove species clapped their
wings together during takeoff and escape flight (clap-and-fling wingbeat described by
Weis-Fogh 1973; Ellington, 1984; Appendix A, Figure Al). The extent to which the
wings overlapped varied within and among species was not measured, but superficially
lacked a noticeable pattern (Appendix A, Figure Al). In addition to the clap-and-fling
wingbeat, several authors have suggested that useful forces are produced during the
wingtip reversal upstroke of doves and pigeons (Brown 1948, 1963; Aldridge 1986;
Norberg 1990; Scholey 1983; Seveyka and Warrick, in prep). Evidence from feather
strains (Corning and Biewener 1998), 3-D kinematic analysis (Warrick and Dial 1998),
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 23
and whole-body accelerations (Seveyka and Warrick in prep; Appendix D) of Rock Doves
suggest that potentially useful forces are produced during the upstroke of a wingtip
reversal wingbeat. In this study, all of the doves studied exhibited a wingtip reversal
upstroke (Figure 5; Appendix A, Figure Al); thus, unless the doves are receiving
significantly different benefits from the use of a wingtip reversal (e.g., larger doves are
producing disproportionally more lift, or receiving disproportionately lower inertial power
requirements), then none of the doves would be expected to have relatively higher force
production or climb power due to this mechanism.
Two other kinematic factors that may vary in a wingbeat cycle are wingbeat
amplitude and the duration of the phases of a wingbeat cycle. If force production is
similar in relative magnitude and duration, and wingbeat amplitude scales independent of
mass, then the scaling of the power available for flight is expected to be largely determined
by wingbeat frequency (Pennycuick 1968, 1969; Lighthill 1977; Weis-Fogh 1977).
Although the relative duration of force production during each wingbeat cycle is difficult
to test directly, the durations of the phases of the wingbeat cycle may be used as to
indicate potential differences between the kinematics of bird species. Among the doves in
this study, the relative durations of the phases of the wingbeat cycle (downstroke, wingtip
reversal phase, and flick phase) did not change with mass (Table 6), which suggests that
all of the species are producing force during the same fraction of the wingbeat cycle.
However, wingbeat amplitude may also vary with mass. Among hummingbirds, wingbeat
amplitude is greater in large hummingbirds during maximal loading flights than in smaller
hummingbirds, and is associated with greater lift production (Chai and Willard 1997).
Although wingbeat amplitude was not measured, the doves in this study all exhibited a
clap-and-fling (maximal excursion) upstroke with a near maximal downstroke (the
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 24
wingtips nearly touching at the bottom of downstroke; Appendix A, Figure Al).
Furthermore, among three species of doves spanning a similar size range as reported here,
wingbeat amplitude does not increase with increasing body mass (Bosdyk and Tobalske
1998). Since the doves in this study appear to be using similar wingbeats (Figure 6, Table
6; Appendix A, Figure Al), there is no strong evidence to suggest that any of the dove
species are experiencing relatively greater lift production or producing lift for a greater
portion of the wingbeat cycle due to their wing kinematics.
Flight Performance: Maximum mass-specific and muscle mass-specific climb power
How do the climb power measurements reported here compare to previously
reported values? In general, the previously reported values for climb power in Rock
Doves appear to be slightly lower than those documented in this study. The mean
maximum mass-specific whole-body climb power in the Rock Doves (29.0 Wkg’*) was
higher than documented in trained Rock Doves carrying a payload and flying vertically
(20.8 Wkg’*; Bosdyk and Tobalske 1998). Mean climb power for the Rock Doves (10.1
W) was also higher than reported for a trained pigeon flying to a perch (8.12 W; Dial and
Biewener 1993). Furthermore, the average pectoralis and supracoracoideus muscle mass-
specific climb power for the Rock Doves was 125.6 Wkg’*, which is higher than the 100
Wkg * documented by Pennycuick and Parker (1966). Differences between the values
reported here and those previously reported for climb power in Rock Doves are probably
due to motivational differences. The doves that were trained to go to a perch (Bozdyk
and Tobalske 1998, and Dial and Biewener 1993) were probably less motivated than the
wild caught Rock Doves used in this study. Furthermore, the wingbeat frequencies of the
captive Rock Doves in Pennycuick and Parker (1966) were mostly under 8.9Hz compared
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 25
to the mean of 9.3Hz reported here in escaping pigeons, also indicating possible
motivational differences.
Among the doves, maximum mass-specific and muscle mass-specific climb power
scaled similar to that predicted by the models of the scaling flight performance
(Pennycuick 1968, 1969, 1989; Lighthill 1977; Weis-Fogh 1977; Ellington 1991; Marden
1994; Figures 7 and 8; Table 9). Assuming muscle strain is independent of mass, which
may not be true for all groups of birds (Tobalske and Dial 1996b, unpublished data),
power available for flight is predicted to be proportional to the combination of force
production, which is proportional to muscle cross-sectional area, and frequency of
contraction (i.e., wingbeat frequency). In this study, body mass-specific climb power was
highest in species with high flight muscle ratios and high wingbeat frequencies (White-
tipped Dove and Inca Dove) and lowest in the species with the lowest flight muscle ratio
and lowest wingbeat frequency (Rock Dove; Table 9, Figure 8). The high flight muscle
ratio of the White-tipped Dove and Inca Dove may have resulted in relatively high mass-
specific force production compared to the other dove species (Pennycuick 1968; Marden
1987). It is important to note, however, that I have made several assumptions in my
estimate of flight muscle mass (e.g., Inca Dove flight muscle ratios are represented by
other members of the genus Columbina, and pectoralis scarring on the keel adequately
reflect the flight muscle ratios of the Mourning and White-winged Doves). Clearly, these
assumptions are not unfounded and are easily testable in the future (similar assumptions
were used by Marden [1990, 1994] and Ellington [1991] to estimate muscle mass-specific
lift and power output for Harris’ Hawks). Although the assumption decreases the
confidence in the slopes of the scaling muscle mass-specific climb power, it is clear that
body mass-specific climb power and muscle mass-specific climb power decreased with
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 26
increasing body mass (Figures 6 and 7, Table 9). The observed decrease in performance
with increasing mass is consistent with the decline in climb power with mass in the
Galliformes (Tobalske and Dial, per s. com), and the decline of acceleration ability with
mass in passeriformes (DeJong 1983) and in aerial insectivores (Warrick 1998).
Furthermore, as predicted from the original models of the scaling of flight performance
(Pennycuick 1968, 1969, 1989; Lighthill 1977; Weis-Fogh 1977), muscle mass-specific
climb power output was highly correlated with wingbeat frequency and negatively
correlated with wingspan and the lengths of the wing bones (Figure 8, Tables 4 and 9).
Most of the variation in mass-specific climb power was explained by a combination
of wing length (also wingspan and wing bone lengths) and flight muscle ratio (Table 10).
Dove species with short wings, had the highest wingbeat frequencies and highest mass
specific climb power (Figures 4 and 8), The strong positive association between flight
muscle ratio and mass-specific climb power is probably due to an increase in available
force production in proportionally larger muscles (Hill 1950; Marden 1987). However,
the observed negative association between wingspan and mass-specific climb power is
more difficult to explain. In theory, relatively long wings should yield reduced induced
power requirements (Pennycuick 1968, 1989; Rayner 1979; Norberg 1990). Yet, a
disproportionate increase in wing length is likely to result in a concomitant increase in the
moment of inertia of the wing and inertial power requirements (Weis-Fogh 1972; Norberg
1990; van den Berg and Rayner 1995). Thus, disproportionate increases in wing length,
wingspan, and lengths of the wing bones among dove species may have resulted in a
disproportionate increase in wing moment of inertia and inertial power costs with
increasing size. Such an increase in moment of inertia of the wing is also expected to
reduce wingbeat frequency (Hill 1950) and the power available for flight (Pennycuick
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 27
1968). However, as stated previously, wingbeat frequency did not decrease as fast as
expected with wing length. Thus, although the doves with relatively long wings had lower
escape flight performance, their wingbeat frequencies and performance were not as low as
would be predicted given their wing lengths.
Data presented here do not allow a direct test of the models on the scaling of flight
performance because lift force, power required, and power available were not measured.
Although wingbeat frequency and muscle size are expected to determine the power
available for flight (Pennycuick 1968), the climb power measurements reported here
cannot be used to determine the scaling of power available or power required for flight for
the doves. Instead, these measurements of Pci represent the difference between the two
types of power, assuming that the power required scales independent of mass. If power
required for flight does scale independent of mass for the doves, then the decline in escape
flight performance may be the result of a decline in the power available for flight from the
decrease in wingbeat frequency with increasing mass. Alternatively, the power required
for flight for the doves may have increased with increasing mass, and with or without a
decrease in power available, would have resulted in decreased flight performance with
increasing size.
Models presented here on the scaling of flight performance (Tables 9 and 10) and
those presented by Marden (1990, 1994) both indicate that flight performance is positively
associated with flight muscle ratio. However, performance is positively associated with
wing length in Marden's model (1994), and negatively associated with wing length in the
models presented here. The multiple regression model for predicting lift to power ratio
from Marden (1994) suggests that lift per unit power output should increase with mass
and increasing wing length among the five species of doves. If the lift to power ratio
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 28
increased with mass, then performance should have also increased with mass unless the
mass-specific power available for flight decreased, or another component of the power
required for flight increased with mass. Thus, the decrease in performance in the doves
would still be expected due to the decrease in power available predicted from the original
models on the scaling of flight performance (Pennycuick 1968, 1969, 1989; Lighthill 1977;
Weis-Fogh 1977).
The muscle mass-specific climb power for the doves in this study span a similar
range as previously reported values for the muscle mass-specific power available for flight
[96 Wkg-1 to 152 Wkg-1, reported here versus 111 Wkg'^ to 177 Wkg'\ reported by
Ellington (1991)]. However, total mass-specific power available for flight should greatly
exceed maximum climb power because total power is the addition of induced, profile,
parasitic, inertial, and climb power. This unexpected similarity suggests that the modes
used by Ellington (1991) and Marden (1994) grossly underestimate the power
requirements. One possible underestimate of power by Ellington (1991) and Marden
(1994) may be the result of assuming that inertial power requirements are low during slow
flight. However, there is evidence that inertial power is important during slow flight. For
example, in a nectar feeding bat inertial power may constitute 60% of the total power
required for flight (Norberg et. al. 1993). Furthermore, van den Berg and Rayner’s (1995)
model for inertial power in birds yields maximum flight muscle mass-specific inertial
power values of 470 Wkg'' for Rock Doves (using 9.3 Hz for wingbeat frequency and
2.79 radians for wingbeat amplitude), which is extremely high for vertebrate muscle
(Weis-Fogh and Alexander 1977; Franklin and Johnson 1997). If the estimate of inertial
power in the Rock Dove is reduced by half, assuming that inertial power is lower on the
upstroke than on the downstroke (van den Berg and Rayner 1995), and then reduced by
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 29
60%, assuming most of the inertial power can be used for aerodynamic power (Norberg
et. al. 1993), then, using my measurements and van den Berg and Rayner’s model (1995),
the muscle mass-specific inertial power requirements for the Rock Dove should be
approximately 94 Wkg'*. Although this value is probably still greatly inflated, it does
support the prediction that inertial power requirements are high during slow flight. Thus,
the results that I present here suggest that the estimates of total power available from
Ellington (1991) and Marden (1994) underestimate some component of power.
The analyses of Marden (1987, 1990, 1994) and Ellington (1991) are dependent
on the assumption that weighing animals until they cannot fly elicits the maximum lift
force that an animal can produce. However, it seems that unless the animals were taking
only one wingbeat, or, rather, one downstroke, that the weight that an animal carried
would not translate directly into maximum lift force. It is unclear from the methods in
Marden (1987) how similarly animals could behave at maximum exertion. Marden (1987)
measured maximum lift at the point between when an animal carrying a weight could
achieve flight and when it could not. If the animals were taking several wingbeats, or
flying vertically or forward at different rates, then their motion was dependent not only on
lift force, but also on the rate at which they produced lift. Unless Marden (1987)
standardized the motion that the animals could produce (i.e., they were all accelerating
similarly or used one wingbeat), then the estimates of lift and estimates of power from the
data (Ellington 1991; Marden 1990, 1994) are inconclusive. If the rate of lift production,
rather than maximum lift force, was measured, then the relationship may have been driven
by wingbeat frequency, resulting in erroneous calculations of power available.
Furthermore, if the animals did accelerate or climb upward then they would have exhibited
higher total power available than assumed based on the lifting measurements of Marden
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 30
(1987).
Alternatively, the relationships between mass and available power reported by
Marden (1990, 1994) and Ellington (1991) for birds may not have been observed if the
birds had been from a closely related group. Differences in wing morphology, muscle
fiber composition, and kinematics may be expected to be higher across distantly related
groups than within a closely related group. Among the birds examined in Marden (1987),
the pigeon used a wingtip reversal upstroke and a clap and fling wingbeat, the
hummingbird used a symmetrical wingbeat, while the other birds probably used
conventional wingbeats. In addition, considerable variation in performance may exist
within a bird species (Table 8). The presence of this type of variation may have severely
influenced the results of Marden (1987, 1990, 1994) and Ellington (1991) because their
sample sizes for bird species were so small (only two of the 10 species studied included
more than one individual).
Lack of phylogenetic control may have also influenced the results of other
investigations into flight morphology and flight behavior. For example, van den Berg and
Rayner (1995) reported values for the scaling of moment of inertia in birds and used these
values to predict the scaling of inertial power for birds. Their moment of inertia
measurements and those of others include small birds such as hummingbirds and sparrows,
and large birds such as Herons and seabirds [Herons wing bone elements do not scale as
predicted by isometry (Appendix F), and seabirds have proportionally heavy wings (van
den Berg and Rayner, 1995)]. It is not clear how these data could be used to make
realistic predictions about the morphology or inertial power requirements of a group, such
as the Columbids, The scaling of wing moment of inertia probably varies between groups
and, perhaps, even within a group that includes ecologically dissimilar birds.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 31
The slopes of the regression lines presented herein were not seriously altered after
re-analysis with the independent contrast method (Felsenstein 1985; Garland et. al. 1992,
using PDTREE, Jones and Garland 1993), probably because phylogeny was already
controlled for by investigating within a group. However, p-values were slightly higher
after the analysis.
In addition to kinematics, morphology, and phylogeny, differences in motivation,
air density, or behavioral responses to the tower may have affected performance in the
doves. However, these factors do not appear to be responsible for the trends reported
here. Motivation levels of the species in this study are unknown; however, escape flight
wingbeat frequency was significantly higher than takeoff wingbeat frequency for each
species, which suggests that the birds were all highly motivated (Table 5). Differences in
air density between the Montana and Texas sites may have negatively affected flight
performance in the Rock Does more than I assumed, yet it seems unlikely that the Rock
Dove would have achieved a mean muscle mass-specific climb power output that would
have been greater than, or equal to, any of the other dove species if the air densities had
been equal. Perhaps the release of doves into a tower with the same dimensions provided
smaller does with more room to fly, and, thus, the chance to achieve higher vertical
velocities and measured power output. In general, however, most of the doves ascended
vertically rapidly after takeoff, and I accounted for most horizontal movement when
determining velocity and acceleration.
Performance versus flight behavior
Foraging and migratory behavior appear to be correlated with escape flight
performance (i.e., mass-specific climb power) among the doves. Mourning Doves, White
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 32
winged Doves, and Rock Doves are all active flyers (Mirarchi and Baskett 1994; Johnston
1992; del Hoyo et.al. 1997; pers.obs.), and all three species had relatively long, high
aspect ratio wings, low wingbeat frequencies and low burst performance compared to the
more sedentary Inca Dove and White-tipped Dove (Figures 7 and 8, Tables 2 and 4;
Mueller 1992; del Hoyo et.al. 1997; pers. obs.). Burst performance in the latter species
appears to be due to their high flight muscle ratios and relatively short wingspan (Table
10); these characteristics may allow the birds to produce high wingbeat frequencies and
muscle power output (Hill 1950; Pennycuick 1968). Nevertheless, escape flight
performance is only one form of flight behavior, and modifications that appear to increase
burst flight performance, probably decrease sustained performance (Pennycuick 1989;
Norberg 1990; Warrick 1998). Carrying extra flight muscle and having short, low-aspect-
ratio wings would undoubtedly yield relatively high costs of transport for long distance or
long term flight performance (Pennycuick 1989; Norberg 1990). Birds that are active,
long distance flyers may benefit, instead, by having relatively small flight muscles and more
efficient long, high-aspect-ratio wings to reduce flight costs (Pennycuick 1968, 1969,
1975, 1989; Epting and Casey 1973; Ellington 1984; Norberg 1990). Adaptations for fast
flight, therefore, probably yield lower burst performance due to relatively smaller flight
muscles, and relatively long wings that reduce force production and wingbeat frequency,
respectively (Warrick 1998).
Although it is unrealistic to make predictions on the maximum size limit of doves
from the data presented here, my models may be easily tested by determining the burst
performance of other dove species that are preferably within the size range studied. For
example, the 140g Grey-chested Dove {Leptotila cassini) may be expected to have high
burst performance. Based on Hartman (1961) the combined pectoralis and
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 33
supracoracoideus flight muscle ratio of L. cassini is 34.2 % (versus 23.5% for the Rock
Dove), which is expected to translate into high force production (Hill 1950; Pennycuick
1968; Marden 1987). Based on my models (Table 10) the predicted mean mass-specific
climb power output for L. cassini is 57.8 Wkg-1 (using 36.4 cm for wingspan, Appendix
A, Table Al).
If burst performance in doves continues to decrease with increasing mass then,
perhaps, large dove species would be more susceptible to predation than smaller species,
unless their size freed them from predation or they occurred in areas that provided lower
predation risk. There may be a relationship between burst performance and the
distribution of some species of birds. It is interesting to note that among the family
Columbidae, most dove species are fairly small (<300 g; Dunning 1995), and many large
dove species (> 450 g) occur on islands. Indeed, the largest species of doves, belonging
to the genus Ducula and the genus Gonra, are restricted mostly to the islands of
Australasia (Baptista et. al. 1997).
Furthermore, flight performance among the Columbids studied here may be
correlated with social behavior. Two species of solitary doves that are rarely found in
flocks, the White-tipped Dove and Grey-chested Dove {Leptotila cassini; Baptista et. al.
1997) appear to be designed for high burst performance; the White-tipped Dove was
found to have high burst performance for its size, while the Grey-chested Dove was
predicted to have relatively high burst performance for its size. In contrast, the highly
social species in this study, the Mourning Dove, White-winged Dove, and Rock Dove,
which often forage, roost, or drink in groups (Baptist et. al. 1997; pers. obs ), had
relatively lower mass-specific climb power for their mass. Perhaps, higher burst
performance in the solitary doves is a response to the potential lack of advanced warning
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 34
of predators that has been predicted to occur in groups of birds (Lima 1987), and has been
supported by vigilance behavior in several species of doves (Burger 1992; Phelan 1987).
An interesting future study into flight performance would be an investigation into the
possible relationship between escape flight performance and the social behavior of birds
species.
Conclusions andfuture considerations
The noticeable decrease in performance documented for doves may be due to an
increase in the power required for flight and/or a decrease in the power available for flight
with increasing mass (Pennycuick 1968, 1969, 1989; Lighthill 1977; Weis-Fogh 1977), or
it may be due to a decrease in the lift produced per unit power with increasing mass
(Ellington 1991, Marden 1994). The data presented here suggested that performance is
negatively associated with increasing wingspan (Table 9 and 10) despite the prediction
that relatively longer wingspans should result in relatively lower mass-specific induced
power requirements (Pennycuick 1968, 1969, 1989; Lighthill 1977; Weis-Fogh 1977;
Rayner 1979). Yet, as expected, mass-specific climb performance was positively
associated with increasing flight muscle mass (Pennycuick 1968).
Future Research
It would be interesting to extend this study to include other large, wild doves, such
as Band-tailed Pigeons {Colwnba fasciatd) or Wood Pigeons {Columba palumbus)^ to
determine if muscle mass-specific climb power continues to decrease in a similar manner
among the Columbidae. It would be of further interest to compare the results presented
here to the scaling of performance and muscle mass-specific climb power output of other
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 35
families of birds. For general comparison, escape flights of two passerine species were
documented. A single Northern Mockingbird {Mimuspolyglottos, 50g, n = 1) had a
maximum mass-specific climb power output of 53 .9 Wkg'\ while the Great Kiskadee
{Pitangiis sulphuratus, 75g, n = 4), a large flycatcher, exhibited a maximum mass-specific
climb power output of 45.5 ± 6.6 Wkg ' (mean ± SD). These values further indicate,
albeit weakly, that mass-specific climb power decreases with increasing mass.
Other species of birds examined, including the Long-billed Thrasher {Toxostoma
longirostre). Yellow-billed Cuckoo {Coccyzus americanus), and Common Ground Dove
{Columbina passerina), did not tend to escape vertically when placed in the tower, but
instead continued to fly into the mesh on the side; thus, the techniques described here may
not be suitable for all species of birds. The species that escaped vertically out of the top of
the tower (e.g. the doves in this study, the Northern Mockingbird and the Great Kiskadee)
fly vertically on a regular basis when flushed or during foraging, while those that did not
fly out of the tower tend to fly forward when flushed (pers. obs). Performance in birds
that tend to flush forward rather than vertically, may be better studied by releasing them
into a “tower” or “tunnel" that was angled slightly or level, similar to Warrick (1998).
REFERENCES
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 36
Aldridge, H. D. J. N., 1986. Kinematics and aerodynamics of the greater horseshoe bat,
Rhinolophus ferrumequinum, in horizontal flight at various flight speeds. J. exp.
Biol. 126:479-497.
Alexander, R, M, 1983. Animal Mechanics. Funtington, Chechester: Packard Publishing
Ltd.
Baptista, L.F., Trail, P.W. and Horblit, H.M. 1997. Family Columbidae (Pigeons and
Doves). Pp. 60-245 in: del Hoyo, J., Elliott, A and Sargatal, J. eds. (1997).
Handbook of the Birds of the World. Vol. 4. Sandgrouse to Cuckoos. Lynx
Edicions, Barcelona.
Bosdyk, A.K. and Tobalske, B.W. 1998. Muscle composition and maximum power for
flight in doves and pigeons (Columbiformes). American Zoologist. 38(5): 126A.
Brown, R.H.J. 1948. The flapping cycle of the pigeon. J. Exp.Biol. 25: 322-333.
Brown, R H J 1963. The flight of birds. Biol. Rev. 38:460-489.
Burger, J. 1992. Drinking, vigilance, and group size in White-tipped Doves and Common
Ground-doves in Costa Rica. Wilson Bulletin 104(2): 357-359.
Chai, P., Millard, D 1997. Flight and size constraints: hovering performance of large
hummingbirds under maximal loading. J. Exp. Biol.; 200(21):2757-2763.
Coming, W. R. and Biewener, A. A. 1998. In vivo strains in pigeon flight feather shafts:
implications for structural design. J. exp. Biol. 201: 3057-3065
Dejong, M.J. 1983. Bounding flight in birds. PhD thesis, University of Wisconsin,
Madison, USA.
del Hoyo, J. Elliott, A, Sargatal, J. eds. 1997. Handbook o f the birds o f the world.
Vol.4. Sandgrouse to Cuckoos. Lynx Edicions, Barcelona.
Dial, K.P. 1992a. Activity patterns of the wing muscles of the pigeon (Columba livia)
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 37
during different modes of flight. J. Exp, Zool. 262:357-373.
Dial, K.P. 1992b. Avian forelimb muscles and nonsteady flight: Can birds fly without
using the muscles in their wings? The Auk. 109(4): 874-885.
Dial, K.P. and A.A. Biewener. 1993. Pectoralis muscle force and power output during
dijfiferent modes of flight in pigeons {Columba livia). J. exp. Biol. 176: 31-54.
Dunning, J B , Jr. (Ed.). 1993. CRC handbook of avian body masses. CRC Press, Boca
Raton, Florida.
Ellington, C.P. 1991. Limitations on animal flight performance. J. exp. Biol. 160:71-91.
Ellington, C P 1984. The aerodynamics of hovering insect flight. Phil, Trans. R. Soc.
Ser. B. 305: 1-181.
Epting, R.J. and Casey, T.M. 1973. Power output and wing disc loading in hovering
hummingbirds. Am. Nat. 107: 761-765.
Felsenstein, J. 1985. Phylogenies and the comparative method. Amer. Nat. 125:1-15.
Franklin, C.E. and Johnston, I.A. 1997. Muscle power output during escape responses in
antarctic fish. J. Exp. Biol. 200: 703-712.
Garland, T. Jr., P H Harvey and A R Ives. 1992. Procedures for the analysis of
comparative data using phylogenetically independent contrasts. Syst. Biol. 41: 18-
32
Goldspink, G. 1977. Mechanics and energetics of muscle in animals of different sizes,
with particular reference to the muscle fibre composition of vertebrate muscle. In
Scale Effects in Animal Locomotion (ed. T J. Pedley), pp. 37-55. New York;
Academic Press.
Greenwalt, C.H. 1962. Dimensional relationships for flying animals. Smithson. Misc.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 38
Coll. 144:1-46.
Greenwalt, C.H. 1975. The flight of birds. Trans. Am. Phil. Soc. 65: 1-67.
Hartman, 1961. Locomotor mechanisms of birds. Smithson, misc. Colins. 143: 1-91
Hill, A. V. 1950. The dimensions of animals and their muscular dynamics. Science
Progress 38: 209-230.
Johnston, R.F. 1992. Rock Dove. In The Birds of North America, No. 13 (A. Poole, P.
Stettenheim, and F. Gills, Eds.) Philadelphia: The Academy of Natural Sciences;
Washington, DC: The American Ornithologists’ Union.
Jones, J A , and Garland, T., Jr. 1993. PDTREEE (version 2.0).
Lighthill, M.J. 1973. On the Weis-Fogh mechanism of lift generation. J. FI. Mech. 60: 1-
17.
Lighthill, M.J. 1977. Introduction to the scaling of aerial locomotion. Pages 365-404 in
Scale effects in animal locomotion (T. J. Pedley, Ed ). Academic Press, New
York.
Lima, S. L. 1987. Vigilance while feeding and its relation to the risk of predation. J.
Theor. Biol. 124(3):303-316.
Marden, J H. 1987. Maximum lift production during takeoff in flying animals, J. Exp.
Biol. 130; 235-258.
Marden, J.H. 1990. Maximum load-lifting and induced power output ofHarris’ hawks are
general functions of flight muscle mass. J. exp. Biol. 149: 511-514.
Marden, J.H. 1994. From damselflies to pterosaurs; How burst and sustainable flight
performance scale with size. Am. J. Physiol. 266: R1077-R1084.
Mirarchi, R E. and Baskett, T.S. 1994. Mourning Dove (Zenaida macroura). In The
Birds of North America, No. 117 (A. Poole, and F. Gills, Eds.) Philadelphia: The
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 39
Academy of Natural Sciences; Washington, DC: The American Ornithologists’
Union.
Mueller, A.J. 1992. Inca Dove. In The Birds of North America, No. 28 (A. Poole, p.
Stettenheim, and F. Gill, Eds. ). Philadelphia: The Academy of Natural Sciences;
Washington, DC: The American Ornithologists’ Union.
Norberg, U.M, 1990. Vertebrate flight: mechanics, physiology, morphology, ecology and
evolution. Springer-Verlag, Berlin.
Norberg, U.M., Kunz, T.H., Steffensen, J.F., Winter, Y. and von Helversen, O. 1993.
The cost of hovering and forward flight in a nectar-feeding bat, Glossophaga
soricaina, estimated from aerodynamic theory. J. exp. Biol. 182: 207-227.
Pennycuick, C.J. and G A. Parker. 1966. Structural limitations on the power output of
pigeon’s flight muscles. J. Exp. Biol. 45:489-498.
Pennycuick, C.J. 1968. Power requirements for horizontal flight in the pigeon Columba
livia. J. Exp. Biol. 49: 527-555.
Pennycuick, C.J. 1969. The mechanics of bird migration. Ibis. Ill: 525-556.
Pennycuick, C.J. 1972. Animal Flight. London: Arnold Press.
Pennycuick, C.J. 1975. Mechanics of flight. IN: Farner, D S , King, J.R. (Eds) Avian
Biology, vol. 5. Academic Press, London, New York, pp. 1-75.
Pennycuick, C.J. 1989. Bird flight performance: a practical calculation manual. Oxford
University Press, Oxford.
Pennycuick, C.J. 1990. Predicting wingbeat frequency and wavelength in birds. J. Exp
Biol. 150: 171-185.
Pennycuick, C.J. 1996. Wingbeat frequency of birds in steady cruising flight: new data
and improved predictions. J. exp. Biol. 199(7): 1613-1618.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 40
Pennycuick, C J , Fuller, M.R., and McAllister, L. 1989. Climbing performance of Harris’
hawks {Parabuteo unicinctus) with added load: implications for muscle
mechanics and for radio tracking. J. exp. Biol. 142: 17-29.
Phelan, J.P. 1987. Some components of flocking behavior in the rock dove (Columba
livia). Journal of Field Ornithology 58(2): 135-143.
Rayner, J.M.V. 1979. A new approach to animal flight mechanics. J. exp, Biol. 80:17-
54.
Rayner, J.M.V. 1985. Linear relations in biomechanics: the statistics of scaling functions.
J. Zool. Lond. 206: 415-439.
Scholey, K.D. 1983. Developments in vertebrate flight. Ph.D. thesis, Univ. Bristol,
United Kingdom.
Sibley, C.G., Ahlquist, J E. 1990. Phylogeny and classification of birds: a study in
molecular evolution. Yale University Press, New Haven, Conn.
Sokal, R.R. and Rohlf, F.J. 1981. Biometry, 2nd ed. New York: W.H. Freeman and Co.
Tobalske, B.W. 1996. Scaling of muscle composition, wing morphology, and
intermittent flight behavior in woodpeckers. Auk 113(1): 151-177.
Tobalske, B.W. and K.P. Dial. 1996a. Flight kinematics of Black-billed Magpies and
pigeons over a wide range of speeds. J. Exp. Biol. 199:263-280.
Tobalske, B.W. and K.P. Dial. 1996b. Scaling of escape-flight performance in galliform
birds. American Zoologist 36(5): 428.
van den Berg, C. and Rayner, J. M.V. 1995. The moment of inertia of bird wings and the
inertial power requirements of flapping flight. J exp. Biol. 198(8): 1655-1664.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 41
Warrick, D.R., and Dial, K.P. 1998. Kinematic, aerodynamic, and anatomical
mechanisms in the slow maneuvering flight of pigeons. J. Exp. Biol. 201: 655-
672.
Warrick, D.R. 1998. The tuming-and liner-maneuvering performance of birds: the cost of
efficiency for coursing insectivores. Can. J. Zool. 76: 1063-1079.
Weis-Fogh, T. and Alexander, R.McN. 1977. The sustained power output from striated
muscle. In Scaling Effects of Animal Locomotion (ed. T.J. Pedley), pp. 511-525.
London: Academic Press
Weis-Fogh, T. 1972. Energetics of hovering flight in hummingbirds and Drosophila. J.
Exp. Biol. 56: 79-104.
Weis-Fogh, T. 1973. Quick estimates of flight fitness in hovering animals, including novel
mechanisms for lift production. J. exp. Biol. 59: 169-230.
Weis-Fogh, T. 1977. Dimensional analysis of hovering flight. In: Pedley T.J. (Ed.) Scale
effects in animal locomotion. Academic Press, London, New York, pp. 405-420.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 42
ACKNOWLEDGMENTS
I thank Michael Delesantro and the Rio Grande Valley Bird Observatory for
providing me with technical assistance, mist nets to tear holes in, and knowledge of local
birds. In addition I thank Mark Conway, and Alan and Joseph Delesantro for helping in
the capture of doves in the brutal Texas heat and humidity. This project would not have
been possible without the generous support and enthusiasm of Ray Keale, an original
nature nut. I also extend my appreciation to many other individuals for helping me along
the way. Ken Dial has patiently provided scientific guidance and support for more years
than either of us care to admit. I cannot begin to explain all of the contributions that Doug
Warrick has made to my thesis, and my development. You were right Doug. Pigeons
rock. Bret Tobalske is in part responsible for my interest in birds, scaling, flight, and
guitar. Because of this, Bret has been subject to a constant stream of questions regarding
each topic, but he always provided excellent answers and great advice. Bret provided
integral insight into this thesis work that resulted in greatly improving the text and
interpretation. Dr. Ebo Uchimoto and Dr. Richard Hutto have contributed to both my
undergraduate and graduate scientific development and their support is appreciated.
Tauzha Grantham helped me stay sane, or at least prevented me from becoming more
insane. Cassie Shigeoka, Ryan Bavis, and Matt Bundle engaged in fruitful discussions on
various aspects of this project. Charla Bitney helped train and collect flight footage of
Rock Doves, in addition to providing names for each pigeon. Generous loans of
specimens or access to specimens were provided by the Phil Wright Zoological Museum,
the Burke Museum, and the Carnegie Museum of Natural History.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Table 1. Morphometric variables for the bones of six species of doves (mean ± SD, with number of birds in parentheses).
Variable Common Ground Inca Dove Mourning Dove White-winged White-tipped Rock Dove Dove Dove Dove Mass (g) 30.1 * 43 .8 * 116.8 : ±14.0 ( 19) 141.9 ± 10. 3 (3) 145.9 ± 11.2 (3) 336.4 ± 29.3 (6)
Sternum Length (mm) ± ± 33.1 ± 1.8 (3) 34.4 0.5 (2) 48.1 1.6 (16) 51.6 ± 0.3 (2) 52.5 ± 2.1 (3) 62.7 ± 4.4 (4) C o Furcula Length (mm) 15.6 ± 0.3(3) 16.1 ± 0.6 (4) 21.7 ±1.0 (14) 25.6 ±1.8 (2) 25.1 ± 1.2 (2) 33.0 ± 0.4 (4) (O
Scapula length (mm) 22.8 ± 0.6 (3) 22.3 ± 1.8 (4) 32.4 ± 1.8 (15) 35.6 ± 0,7 (2) 37.4 ± 1.5 (3) 43.5 ± 2.5 (5) &
Coracoid length (mm) 18.1 ± 0.3 (2) 18.4 ± 0.7 (4) 25.5 ± 2.1 (19) 27.8 ± 0.7 (2) 28.3 ± 0.6 (3) 36.6 ± 1.3 (5)
Humerus length (mm) 20.8 ± 0.5 (3) 20.8 ± 0.6 (4) 31.6 ± 0.9 (18) 35.3 ± 1.5 (3) 34.0 ± 1.5 (3) 46.1 ± 3,1 (6) % Radius length (mm) 22.0 ± 0.9 (3) 22.2 ± 0.3 (3) 32.8 ± 1.0 (19) 37.2 ± 0.4 (3) 34.4 ± 1.1 (3) 49.9 ± 2,5 (6) 2 Ulna length (mm) Û. 24.5 ± 0.7 (3) 25.3 + 0.7 (3) 36.4 ±1.1 (19) 42.6 (1) 38.4 ± 1.3 (3) 55.1 ± 3.0 (5) 0c Ü Manus length (mm) 14.5 + 0.4 (3) 14.4 ± 0.2 (4) 22.9 ± 0.8 (19) 25.6 ± 0.3 (3) 23.6 ± 1.3 (3) 34.5 ± 1.6 (6) 3 1 Û. Pectoralis Scar Area (cm^) 0.8 ± 0,0 (3) 0.9 ± 0.1 (3) 2.1 ± 0.3 (17) 2.4 ± 0.4 (2) 2.5 ;t (3.4 (3) 3.8 ± 0.4 (4) P * Mass from Dunning (1993)
CCD o g 8 CD O c g
CD Û.
8 3 "O 2 Q. CD q: 5
Table 2. Least-squares and Reduced-major-axis (RMA) regression slopes, and correlation coefficients (r) describing relationships between log of morphometric variables and log body mass for 26 species of doves and the 6 species of doves in this study. 26 species 6 species
Variable Least-squares RMA r Least-squares RMA r
Total Wing Area 0.72 0.75 0.97* 0.73 0.73 LOO* Wingspan 0.40 0.42 0.96* 0.40 0.41 0.98* (/>CO
CD Winglength 0.39 0.40 0.96* 0.40 0.41 0.98* Q. Wing Loading 0.28 0.34 0.81* 0.26 0.27 0.97* Wing Disk Loading 0.19 0.30 0.65* 0.19 0.25 0.75 "O CD Aspect Ratio 0.07 0.12 0.60* 0.08 0.14 0.58 2 Pectoralis Scar Area 0.63 0.69 0.92* 0.72 0.73 0.99* Q. Ulna Length 0.40 0.40 0.98* 0.36 0.36 0.99* O "O3 2 Sternum Length 0.29 0.31 0.95* 0.29 0.30 0.99* Q. CD
Radius Length 0.40 0.40 0.98* 0.36 0.37 0.99* CD t: Humerus Length 0.39 0.40 0.99* 0.36 0.37 0.99* Manus Length 0.39 0.40 0.99* 0.39 0.40 0.99* Coracoid Length 0.35 0.35 0.99* 0.32 0.32 0.99* g Scapula Length 0.30 0.30 0.98* 0.31 0.32 0.98* 8 Furcula Length 0.35 0.36 0.99* 0.33 0.34 0.99* O * p < 0.05 c o '
CD Q.
"O 8 "O3 2 Q. Ù1CD trt
Table 3. Morphometric Variables for the wings of six dove species (mean ± SD, with number of birds in
Variable Common Ground Inca Dove Mourning Dove White-winged White-tipped Dove Rock Dove Dove Dove Mass (g) 30.1 * 48.1 ± 3.7 (22) 105.2 ±13.0 (5) 146.2 ± 19.3 (6) 172.5 ± 12.2 (14) 349.6 ± 39.1 (28)
Wing Area (cm^) 56.4 ± 4 .0 (3) 64.5 ± 3.9 (23) 127.8 ± 6.9 (5) 156.0 ± 20.6 (6) 156.8 ± 7.4 (14) 311.6 ± 24.6 (28)
Wing Length (cm) 11.7 ± 0.3 (3) 12.2 ± 0.7 (23) 19.1 ± 0.4 (5) 21.0 ± 1.3 (6) 18.6 ± 0.7 (14) 30.5 ± 1.2 (28) (/>(/)
Aspect Ratio Index 4.8 ± 0.1 (3) 4.6 ± 0.4 (23) 5.7 ± 0.4 (5) 5.7 ± 0.2 (6) 4.4 ± 0.2 (14) CD 6.0 ± 0.3 (28) Q. Base Wing Chord (cm) 6.3 ± 0.3 (3) 6.7 ± 0.2 (23) 8.7 ± 0.3 (5) 9.9 ± 0.4 (6) 10.7 ± 0.6 (14) 12.0 ± 0.6 (28)
Body Width (cm) 2.7 ** 2.8 ± 0.3 (23) 3.6 ± 0.1 (5) 3,7 ± 0.4 (6) 4.8 ± 0.3 (14) 7.3 ± 0.8 (28) "O CD Total Wingarea (cm^) 130.0 ± 8.7 (3) 147.6 ± 7.6 (23) 286.7 ± 16.7 (5) 343.9 ± 45.5 (6) 365.7 ± 17.3 (14) 710.2 ± 57.2 (28) 2 Q. Wing Loading (Nm'^) 26.5 ± 1.7 (3) 32.0 ± 3.1 (22) 37.6 ± 5.8 (5) 41.8 ± 2.4 (6) 46.5 ± 3.5 (14) 48.4 ± 5.0 (28) gC Wing disk loading (Nm'^) "G 6.4 ± 0.3 (3) 7.8 ± 2.0 (23) 7.9 ± 1.2 (5) 8.8 ± 0.3 (6) 12.2 ± 1.2 (14) 9.4 ± 1.1 (28) "O3 2 Q. Wing Span (cm) 26.1 ± 0.7 (3) 27.3 ± 1.4 (24) 41.8 ± 0.9 (5) 45.7 ± 2.7 (6) 42.1 ± 1.4 (14) 68.2 ± 2.7 (28) 2 Aspect Ratio 5.2 + 0.1 (3) 5.0 ± 0.3 (23) 6.1 ± 0.4 (5) 6.0 ± 0.2 (6) 4.9 ± 0.2 (14) 6.6 ± 0.3 (28) ■c
CD C I £ g È 8 gCD O C COg CO
CD Q.
"O 8 "O3 2 Q. Q1CD "4-VO
Table 4. Least-squares (and 95% confidence intervals), Reduced-major-axis (RMA) slopes and correlation coefficients (r) describing the relationship between log of flight variable and log wingbeat frequency for 5 or 6 species of doves
Dependant Variable Independant Variable Least -squares RMA r P Level flight wingbeat frequency (6spp.) Body Mass -0.28 (± 0.23) -0.32 0.87 0.029 Take-off wingbeat frequency (6spp.) Body Mass -0.30 (± 0.09) -0.30 0.99 0.001 (/) Escape Flight wingbeat frequency (5spp.) Body Mass -0.34 (± 0.17) -0.35 0.97 0.007 (/> 0) Wing Span -0.78 (±0.14) -0.78 0.995 < 0.001 Q. Pec and Supra FMR 1.51 (±3.94) 2.62 0.57 0.31 Pec and Supra Mass -0.37 (± 0.12) -0.38 0.98 0.002 "O CD Humerus Length -0.88 (± 0.21) -0.89 0.99 0.001 2 Ulna Length -0.89 (± 0.24) -0.90 0.99 0.001 Q. oC Radius Length -0.87 (±0.16) -0.88 0.99 < 0.001 "G "O3 Manus Length -0.81 (±0.12) -0.81 1.00 < 0.001 2 2Q. Pectoralis Scar Area -0.47 (± 0.18) -0.48 0.98 0.004 Total Wing Area -0.44 (± 0.15) -0.45 0.98 0.002 ■c
CD
8
O c g '(/)(/>
CD Q.
"O 8 "O3 2 Q. Q1CD Table 5. Wingbeat frequecies (Hz) for six species of doves (mean ± SD, with number of flights in parentheses)
Species n Take-off Wingbeat n Tower Wingbeat n Level flight Wingbeat Frequency Frequency Frequency Common Ground Dove 4 17.5 ± 1.2 (23) - - 3 10.8 ± 1.4 (7) Inca Dove 4 16.4+1.3 (32) 16 19.0+1.2(16)* 2 13.1 ± 3 .0 (3) Mourning Dove 6 11.8 ± 0.8 (18) 5 12.9 + 0.9(5) * 5 8.7 ± 0.9 (25) (/>(/)
White-winged Dove 6 11.0+1.2(14) 4 12.4 ± 0.8 (4) * 5 6.9 ± 1.0 (36) CD Q. White-tipped Dove 2 10.8 ± 1.1 (4) 6 13.0 ± 1.1 (6) * 2 9.3 ± 0.9 (4) Rock Dove 24 8.9 ± 0.4 (24) 14 9.3 ± 0.4 (14)* 7 6.2 ± 1.1 (14) "O CD
* significantly different than take-off wingbeat frequency (p < .05) 2 Q. gC "G "O3 2 2Q. t:
g 8
O g '(/)(/>
CD Q.
"O 8 733 2 Q. Q1CD 00
Table 6. The percentage (mean ± SD) of the wingbeat cycle that is downstroke, wing-tip reversal, or flick phase in: Common Ground Doves, Inca Doves, Mourning Doves, White-winged Doves, White-tipped Doves, and Rock Doves.
Species n Downstroke Wing-tip Reversal Flick
Common Ground Dove 2 (3) 48.7 ± 2.2 28.1 ± 2.9 23.2 ± 1.8 (/) Inca Dove 4 (28) 52.3 ± 3.9 24.0 ± 3.8 23.7 ± 3.0 (/>
Mourning Dove CD 2 (5) 52.1 ± 4.5 28.9 ± 4.9 19.0 ± 3.2 Q. White-winged Dove 4 (27) 52.8 ± 2.1 25.3 ± 2.0 22.0 ± 2.6 White-tipped Dove "O 4 (28) 52.7 ± 4.2 26.4 ± 3.9 21.0 ± 3.3 CD Rock Dove 3(1 1 ) 52.1 ± 3.7 27.2 ± 3.7 20.6 ± 2.6 2 Q. gC "G "O3 2 Q.2 ■c
g 8
O c g '(/)CO
CD Q.
"O 8 "O3 2 Q. Q1CD (/>(/)
CD Q. o\ Tl- "O CD Table 7. Average Velocity and Acceleration of Doves Escaping from the Tower (mean ± SE) 2 Q. Species n Average Velocity (m/s) Average Acceleration to Peak Velocity (m/s'^2) Cg "G3 Inca Dove 3.4 ± 0.2 4.2 ± 0.5 "O 16 2 Q. Mourning Dove 6 3.1 ±0.3 3.8 ± 0.3 2 White-winged Dove 6 3.0 ± 0.2 3.6 ±0.5 ■c White-tipped Dove 5 3.3 ± 0.3 4.8 ± 0.4 Rock Dove 11 2.4 ± 0.2 1.6 ±0.2
g 8
O gCO CO
CD Q.
"O 8 "O3 2 Q. Q1CD o >n
(/) Table 8. Maximum mass-specific, and muscle mass-specific, whoîe-body power output for 5 species of doves (mean ± SE) (/>
Species n Mass Mass- Pectoralis Mass- Pectoralis and Total Flight CD Q. specific specific Supracoracoideus Muscle Climb Climb Power Mass-specific Climb Mass-specific Power Power Climb Power "O Inca Dove t 16 47.4 ± 1.0 50.5 ± 3.0 228.2 ± 12.7 185.0 ± 10.3 152.4 ± 8.5 CD 2 Mourning Dove 0 6 111.7 ±5.9 45.2 ± 4.3 192.2 ± 18J /55.g 129.1 ±12.3 Q. gC White-winged Dove 0 6 150.0 ± 10.0 44.1 ±3.4 179.9 ±20.2 145.8 ± 16.4 120.8 ± 13.6 "G "O3 White-tipped Dove * 6 174.5 ±4.3 47.1 ±5.0 190.8 ±20.2 157.1 ± 16.7 127.7 ± 13.5 2 2Q. Rock Dove * 11 29.5 ± 1.0' 147.6 ± 4.9 125.6 ±4.2 95.2 ± 3.2 ■c * Muscle mass percentages of total body mass are from Hartman (1961); White-tipped Dove: pectoralis = 24.7%, pectoralis 3 and supracoracoideus = 30%, flight muscle mass = 36.9%; Rock Dove: pectoralis = 20%, pectoralis and supracoracoideus = 23.5%, flight muscle mass = 31% t Muscle mass percentages of total body mass are average measurements of dove species from the genus Columbina from Hartman (1961); pectoralis = 22.7%, pectoralis and supracoracoideus = 28%, flight muscle mass = 34%.
0 Estimated flight muscle masses The values used were: pectoralis = 23.5%; pectoralis and supracoracoideus = 29%; flight muscle mass = 35%. O C • Mass-specific power output was not significantly different between the Inca, Mourning, White-winged and White-tipped g Doves; however, mass-specific power output of the Rock Dove was significantly lower than any of the other dove species. '(/)(/>
(D Q.
"O 8 "O3 2 Q. Q1CD «r>
Table 9. Regression slopes (Least-squares and Reduced-major-axis), regression constants, andcorrelation coefficients describing relationships between log of variable and log mass-specific power output or log muscle mass-specific power output for 5 species of doves (/>(/)
CD Mass-Specific Climb Power Muscle Massespecific Climb Power Q. Variable Least- RMA A r P Least- RMA A r P squares squares "O CD Mass -0.241 0.290 2.14 0.832 0.081 -0.183 -0.192 2.58 0.955 0.011 2 Pectoralis & Supracoracoideus FMR 1.980 2.397 -1.23 0.826 0.033 - - - - - Q. gC Pectoralis Scar Area -0.304 -0.399 1.73 0.761 0.135 - -- -- "G "O3 Pectoralis & Supracoracoideus Mass -0.246 -0.312 2.01 0.788 0.113 - - - - - 2 2Q. Escape Flight Wingbeat Frequency 0.718 0.832 0.83 0.863 0.060 0.540 0.550 1.58 0.990 0.002 .c■C Wing Disk Loading -0.138 -1.160 1.76 0.119 0.849 -0.219 -0.771 2.39 0.284 0.643 Wing Loading -0.844 -1.262 2.99 0.669 0.216 -0.702 -0.834 3.32 0.842 0.074 Wingspan -0.586 -0.650 2.59 0.902 0.036 -0.427 -0.429 2.88 0.995 < 0.001 Wing Length -0.578 -0.646 2.37 0.895 0.040 -0.426 -0.427 2.73 0.997 < 0.001 Aspect Ratio -1.245 -1.613 2.57 0.772 0.126 -0.860 -1.068 2.83 0.805 0.100 Humerus Length -0.600 -0.734 1.34 0.817 0.092 -0.474 -0.486 1.95 0.975 0.005 O Ulna Length -0.639 -0.750 1.36 0.852 0.067 -0.490 -0.500 1.98 0.990 0.001 C g Radius Length -0.625 -0.728 1.34 0.859 0.062 -0.477 -0.481 1.96 0.991 0.001 '(/)(/> 0.077 -0.439 -0.445 Manus Length -0.564 -0.673 1.27 0.838 1.91 0.986 0.002 CD Q.
"O 8 "O3 2 Q. Q1CD (/) (N (/> «n CD Q. O3 Table 10. Multiple regression models describing the relationships between the Log of two variables and Log Mass-specific Climb Power (Log MSCP) "O CD
2 B (Log y) C (Log z) A H P Q. -0.352 (Log L) 1.222 (Log FMRps A) 0.440 1.000 <0.001 Cg "G -0.346 (Log Iw) 1.248 (Log FMRps A) 0.273 0.999 0.001 "O3 2 -0.389 (Log L) 1.041 (Log FMRps B) 0.756 0.994 0.006 Q. 2 -1.514 (Log L) 0.594 (Log Pscar) 3.911 0.975 0.025 -0.503 (LogW) 1.674 (Log FMRps A) 0.021 0.966 0.034 ■c -0.344 (Log L) 1.241 (Log FMRps C) 0.415 0.948 0.052
Log (MSCP) = A + B (Log Y) + C ( Log z). L, Wingspan; Iw, wing length; FMRps, combined pectoralis and supracoracoideus flight muscle ratio [Assumed to be (A) 29%, (B) 30%, or (C) 27% for g Mourning Dove and White-winged Dove]; Pscar, pectoralis scar area on the keel; 8 W, wing loading
(/>(/)
CD Q.
"O 8 "O3 2 Q. Q1CD 53
FIGURE LEGENDS
Figure 1. Captured doves were released into a 2.5m X Im^ tower, and allowed to escape out of
the top. Escape flights were recorded by two cameras; one camera (approximately 8m from the base of the tower) recorded the entire flight and a second camera (approximately 3m from the tower) recorded the kinematics of the bird. Figure 2. Least-squares regression lines, RMA regressions, and correlation coefficients (r) describing the relationships between selected log transformed variables and log body mass for: Common Ground Dove; Inca Dove; Mourning Dove; White-winged Dove; White-
tipped Dove; and Rock Dove. (A) the area of pectoralis scarring on the keel (cm^); (B) ulna
length (cm); (C) manus length (cm); (D) total wing area (cm^); (E) wingspan (cm); (F)
aspect ratio; (G) wing loading (Nm'^); (H) wing disk loading (Nm'^)_
Figure 3. Least-squares regression lines, RMA and correlation coefficients (r) describing the relationships between log wingbeat frequency and log body mass for: Common Ground Dove, Inca Dove, Mourning Dove, White-winged Dove, White-tipped Dove, and Rock Dove. (A) takeoff wingbeat frequency (Hz); (B) escape flight wingbeat frequency (Hz); (C) level flight wingbeat frequency (Hz). (Note: the Common Ground Dove is not included in the escape flight comparison)
Figure 4. Least-squared regression lines, RMA regressions, and correlations coefficients (r)
describing the relationships between selected log transformed variables and log wingbeat frequency (takeoff or escape flight) for: Common Ground Dove, Inca Dove, Mourning
Dove, White-winged Dove, White-tipped Dove, and Rock Dove. (A) log ulna length (cm)
and log takeoff wingbeat frequency (Hz); (B) log ulna length (cm) and log escape flight
wingbeat frequency (Hz); (C) log wingspan (cm) and log takeoff wingbeat frequency (Hz);
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 54
(D) log wingspan (cm) and log escape flight wingbeat frequency (Hz). Figure 5. Examples of lateral views of takeoff flights illustrating the path of the wingtip and wrist with respect to the body during a characteristic takeoff wingbeat of an Inca Dove, Mourning Dove, and White-winged Dove. The bird silhouettes illustrate the body position at the top of upstroke/ beginning of downstroke. The downstroke is illustrated with open circles and the upstroke is illustrated with shaded circles, (note: figures are not to scale; video taken at 250 Hz for four smallest species and 500 Hz for the Rock Dove). Figure 6. Least-squared and RMA regressions and correlation coefficients (r) describing the relationship between log maximum mass-specific whole-body climb power output and log body mass for the doves studied. Figure 7. Least-squares and RMA regressions and correlation coefficients describing the relationship between log body mass and (A) log pectoralis mass-specific climb power, (B) combined pectoralis and supracoracoideus climb power, and (C) flight muscle mass- specific climb power. Figure 8. Least-squares regression lines, RMA regressions, and correlation coefficients (r) describing the relationship between selected log transformed variables and log maximum combined pectoralis and supracoracoideus mass-specific climb power for: Inca Dove, Mourning Dove, White-winged Dove, White-tipped Dove, and Rock Dove. (A) log wing
disk loading (Nm‘^); (B) log escape flight wingbeat frequency (Hz); (C ) log wingspan
(cm); (D) log ulna length (cm); (E) log wing loading (Nm'^); and (F) log aspect ratio.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 55 Figure 1
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Figure 2 56 0.6 # Rock Dove 3.98 y = Q.724X - 1.202 = 0.982 (N r = 0.991, p < 0.001; RMA = 0.73 es White-tipped Dove White-winged Dove^. -2.51
Mourning Dove *
< 0.2- -1.58
0.0- - 1.00 - # Inca Dove Pectoralis Scar Area (5Û Common Ground Dove O h
- 0.2 0.63 2.0 2.5 ( 100) (316) Log Body Mass (g) (Body Mass [g])
1.8 63.1 y = 0 J58x 0.831 r * = 0.980 r = 0.990, p < 0.001; RMA = 0.362 Rock Dove
1.7- -50.1
Whiie-winged Dove# -C ■So 1-6- -39. # White-tipped Dove ■BbO ® Mourning Dove
^ 1.5- -31.6 DO
1.4- # Inca Dove -25.1 Common Ground Dove Ulna Length
20.0 2.0 2.5 (3L6) (100) (316) Log Body Mass (g) (Body Mass [g])
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 57 Figure 2 continued
1.6 r39.8 y = 0391X + 0.546 = 0.978 r = 0.989, p < 0.005; RMA = 0.395 Rock Dove
1,5- -31.6 e
White-winged Dove %
• White-tipped Dove Mourning Dove
- 20.0 ^
1.2- Manus Length -15.8 Inca Dove Common Ground Dove 12.5 2.0 2.5 3.0 (31.6) (100) (316) ( 1000) Log Body Mass (g) (Body Mass [g])
3.0 1000 y = 0.73Ix + 0.963 r : = 0.992 r = 0.996, p < 0.001; RMA = 0.734 Rock Dove # 2.8- -631 CN
P 2.6- -398 W hite-w ingcJ Dove # White-tipped Dove tSO
2.4- -251 L.
Total Wing Area 2.2- -159 Inca Dove
lommon Ground Dove 2.04— 100 1.5 2.0 2.5 (31.6) (100) (316) Log Body Mass (g) (Body Mass [g])
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Figure 2 continued 58 1.9 y = 0.404% + 0.775 r * = 0,958 r = 0.979, p < 0.005; RMA = 0.413 Rock dove #
-63
-50 B White-winged Dove § 0 White-tipped Dove cm c -40
Wingspan -32
Inca Dove ‘ommon Ground Dove 2.5 (316) Log Body Mass (g) (Body Mass [g])
0.85 y = 0.078% + 0.585 r*= 0.331 r = 0.575, p > 0.15; RMA = 0.135
• Rock Dove
0.80- -6.3 # Mourning Dove
g* 0.75 - hite-winged Dove -5.6 < bû Ic/3 Common Ground Dove <
# Inca Dove 0.70- -5,0 # White-tipped Dove
Aspect Ratio 0.65 4.5 1.5 2.0 2.5 (31.6) (100) (316) Log Body Mass (g) (Body Mass [g])
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Figure 2 continued. 59 1.7 •50.1 y = 0.256% + 1.055 r ' = 0.933 • Rock Dove r = 0.966, p < 0.01; RMA = 0.265 White-tipped Dove •
rd • White-winged Dove Ê & Mourning Dove 60 e
60 U 60 Inca Dove 0 a 1.5- -31.6 60
Wing Loading
# Common Ground Dove
1.4 ■25.1 1.5 2.0 2.5 (31.6) (ick)) (316) Log Body Mass (g) (Body Mass [g]>
12.6 1.1 • white-tipped Dove y = 0.193x -t- 0.546 r ^ = 0.566 r = 0.752, p > 0.08; RMA = 0.257
fN 1.0- # Rock Dove White-winged Dove
^0.9- Inca Dove # # .Mourning Dove
Wing Disk Loading # Common Ground Dove 0.8 6.3 1.5 2.0 2.5 (31.6) (100) (316) Log Body Mass (g) (Body Mass [g])
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. CD ■ D O.O C g Q.
■D CD
C/) C/) 1.3 r20.0 y = -0.288% + 1.688 r 0.957 r = 0.978, p < 0.01; RMA = -0.294 8 Common Ground Dove
(O' Inca Dove -15.8 a & g 3. s 3" CD ? ■DCD O - 12.6 Q. s C a 0 White-lipped Dove « 3O Mourning Dove # "O O White-winged Dove #
CD (U Q. - 10.0 g ■D CD Rock Dove C/) C/)
0.9 7.9 1.5 2.0 2.5 (31.6) (100.0) (316.2) Log Body Mass (g) (Body Mass [g]> CD ■ D O Q. C g Q.
■D CD W
C/) 1.3 ■ 20.0 c/) 31 # Inca Dove y = -0.336% + 1.830 r% = 0.945 1 w r = 0.972, p < 0.01; RMA = -0.346 8 a ci' I I* 1.2 J -15.8 a O) I 3 cr k g k C/) C/) 0.9 T T .7.9 1.5 2.0 2.5 (31.6) (100.0) (316.2) Log Body Mass (g) (Body Mass [g]) o\ CD ■ D O Q. C g Q. ■D CD n C/) C/) 1.2 y = -0.276X + 1.510 = 0.749 3 r = 0.865, p < 0.01; RMA = 0.319 1 8 Inca Dove OJ ci' 0 I - 12.6 § 1 a g 3. p 3" Common Ground Dove O' CD - 10.0 While-lipped Dove ■DCD IM O C Q. C a Mourning Dove 3O "O O 0) CD > Q. While-winged Dove # two o Streptopelia decaocto > ■D J 0.8 -6.3 CD Rock Dove* ~ C/) C/) Columba palumbiis Ô 0.7 5.0 1.5 2.0 2.5 (31.6) (100.0) (316.2) Log Body Mass (g) O s (Body Mass [g]) to Figure 4 63 1.3 20.0 i Common Ground Dove I 1-2Inca Dove 15.8 ? g- s £ s I 1.1 . 12.6 f U) ...«While-tipped Dove e Mourning Dove • 2 ^s.^White-winged Dove “ 1 1.0- - 10.0 e2 ^>s^Rock Dove ■I y= -0.811X + 2.358 r'= 0.990 I r = 0.995, p < 0.001; RJVIA = 0.815 0.9 7.9 1.4 1.6 1.8 (25.1) (39.8) (63.1) Log Ulna Length (cm) (Ulna Length [cm]) B 20.0 Jnca Dove S -15.8 2" Mourning Dove • ^White-tipped Dove I 4 c •White-winged Dove - 12.6 I 1.0 - - 10.0 •ock Dove y = -0.892X + 2.526 r^= 0.980 Î r = 0.990, p < 0.001; RMA = -0.901 0.9 7.9 1.4 1.6 1.8 (25.1) (39.8) (63.1) Log Ulna Length (cm) (Ulna Length [cm]) Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Figure 4 continued 64 1.3 20.0 y = -0.709X + 2.232 r*= 0.985 Common Ground Dove r = 0.992, p < 0.001; RMA = -0.715 Inca Dove S 1.2 -15.8 - 12.6 White-tipped Dove Mourning Dove White-winged Dove* is ^ 1.0 - 1 0 .0 O * Rock Dove 0.9 7.9 1.4 1.6 1.8 2.0 (25.0) (40.0) (63.0) (100.0) Log Wingspan (cm) (Wingspan [cm]) D 20.0 Inca Dove -15.8 I I I White-tipped Dove f i , . Mourning Dove '^White-winged Dove - 12.6 I Jo ,E §•1.0 ■ 10.0 Q. I jlock Dove s -0.777% 2.383 r- =0.990 S + w 0.995, p < 0.001: RMA = -0.779 0.9 7.9 1.4 1.6 1.8 2.0 (25) (40) (63) (100) Log Wingspan (cm) (Wingspan [cm]) Reprotducetd with permission of the copyright owner. Further reprotduction prohibitetd without permission. Figure 5 65 Inca Dove Mourning Dove Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Figure 5 continued 6 6 White-winged Dove / White-tipped Dove Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 67 Figure 5. Continued. f Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. CD ■ D O Q. C g Q. 1.8 63.1 3 ■D CD 1 « ON C/) C/) 8 1.7- 50.1 Inca Dove ci' £ ^White-tipped Dove X! £ S Mourning Dove B u *White-wingedDove 3 ï^i U 3" 'O CD 0 X 1.6 - 39.8 -t3 ■DCD 1 0 O (U X1 Q. C o 0) a X: O 3O X "O O CJ c V CD 1 Q. a % CA 1.5- -31.6 a* g y = .0.241% + 2.143 r ' = 0.693, p = 0.081 Vi CQ ■D # Rock Dove CD r = 0.832, RMA = -0.290 i C/) C/) g Phylogenetically corrected r2 = 0.734, p = 0.143, RMA = -0.289 1 4 25.1 1.5 2.0 2.5 3.0 (31.6) (100.0) (316.2) (1000.0) 00o\ Log Body Mass (g) (Body Mass [g]) CD ■D O Q. C g Q. ■D 2.4 251.2 CD 31 C/) C/) 1 Inca Dove • > CD 8 - 200.0 P4 U Dove # While-lipped DoveMourning 3. V 3" CD u Will le-winged Dove • ■DCD o> O a Q. w C I a W]l/J 3O CS "O è O t 2.2- -158.5 CD Q. 2 S Rock Dove • ■D CD y = -0.206X + 2.710 r^ = 0.924 C/) r - 0.961, RMA = -0.214 C/) 125.9 2.0 2.5 (31.6) (100,0) (316.2) ON Log Body Mass (g) NO (Body Mass [g]) CD ■D O Q. C ag . ■D 2 .3 1 9 9 .5 CD 21 c/) c/) 3 Inca DovelK. 1 'ë •-J i Cd 8 ' d 8 2 ci' 8 2 f i g 2.2 158.5 ^ a u • White-tipped Dove 3 Mourning Dove#N. CA ■ s fi & y = -0.183X + 2.575 = 0.911 : r CD •PÜN Q. . f i É r = 0.955, RMA = -0.192 I i . f i 5 1 Phylogenetically Corrected ■D V I CD OA Least squares = -0.208, r = 0.936, O u C/) C/) p = 0.064, RMA = -0.223 2.0 100.0 1 .5 2.0 2 .5 (3 1 .6 ) (100.0) (3 1 6 .2 ) Log Body Mass (g) (Body Mass [g]) o CD ■ D O Q. C g Q. ■D 2.2 158.5 CD Inca Dove • Î C/) C/) wi I n•vj 8 ci' 0 Mourning Dove # White-tipped Dove 125.9 s Whitc-wiiiged Dove 3 u 3" u CD ■DCD O 1 Q. A aC ? 3O % "O cd O - 100.0 CD Rock Dove # Q. I ■D % CD y = -0.221x + 2.566 r ^ = 0.907 W) C/) C/) r = 0.953, RMA = -0.232 1.9 - - 79.4 1.5 2.0 2.5 (31.6) (100.0) (316.2) Log Body Mass (g) (Body Mass [g]) Figure 8 72 2J 1993 Vi 3 « 1— 1 ' o A g s CS u k l Mourning Dove# White-tipped Dove # ’1583 a , 3 © e u White-winged Dove* - 3 S JS CS B u • s a CS Ü k , ca S o C ' 3 Rock Dove # -125.9 s . A I I cn ■ o Z g w t/3 e K CS II Z s E © y= .0.235%+ 2.411 r2 = 0.081 u r = 0.285, RMA = -0.826 2.0 100.0 0.9 1.0 (7.9) (10.0) (12.6) Log Wing DiskD isk Loading (Nm-2) (Wing Disk Loading [Nm-2]) Cfl 2 3 ■199.5 s y = 0.543% + 1.580 r := 0.972 0» 8 r = 0.986, RMA = 0.592 2 Inca Dove o b£ Ï a. White-tipped Dove s 2.2 - t/3 -158.5 Î 'O I Moiirninu: Dove ^ % O s a TS -Q Vi • White-winged Dove CS S 6 Vi U U ts C I0> 8 1 • Rock Dove 425.9 "3a ? -O ? Z B i g S 2 1O ^ eo u 2.0 100.0 0.9 1.0 1.2 1.3 (7.9) (10.0) (15.8) (20.0) Log Tower (escape) Wingbcat Frequency (Hz) (Tower Wingbeat Frequency [Hz]) Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Figure 8 continued 73 2J 199.5 ■a Inca Dove il White-tipped Dove -158.5 « 11 Mourning Dove II White-winged Dove # -a S Rock Dove -125.9 I I"- -O ? VI II .a w y = -0.427X -H 2.882 r - = 0.987 r = 0.994 , R M A = -0.430 2.0 100.0 1.6 1.8 2.0 (39.8) (63.1) (100.0) Lo; (cm ) 23 199.5 Inca Dove ■o ^(White-lipped Dove .158.5 5 5 1 “ Mourning Dove White-winged Dove 11 O 0 11 Rock Dove # -125.9 u ■; I ^ I s y = -0.492% -h 2.962 r - = 0.983 oc r= 0.991, RMA = -0.496 2.0 - — 100.0 1.4 1.5 1.7 1.8 (25.1) (31.6) (39.8) (50.1) (63.1) Log Ulna Length (log cm) (Ulna Length [cm]) Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Figure 8 continued E 74 2.3 199.5 Inca Dove li M ourning Dove • White-tipped Dove • -158.5 2 White-winged Dove# II 1 1 Rock D ove# -125.9 ll y = -0.687x + 3.287 r : = 0.678 r = 0.823, RMA = -0.834 2.0 -1- 100.0 1.5 1.7 (31.6) (50.1) Log Wing Loading (Nm*2) (Wing Loading [Nm*^]) ■199.5 0^ Vi 3 Oi 2 "5 I w *0J3 (9 O 1 £ CO White-tipped Dove # •158.3 A # Mourning Dove k •a ta. B 3 > ot % o Vi Whiie-wineed Dove# * 3 Pm C Xl 3 s 1 .25 CJ 2 □ u u £ o ts "3 Rock Dove # 125.9 a 1 2 "3 S V V3 eg Ô 2 S y = -0.838% + 2.812 = 0.575 o s r = 0.758, RMA = -1.106 V 2.0 100.0 0.6 0.7 0.8 0.9 (4.0) (5.0) (6.3) (7.9) Log .A-speet Ratio (.Aspect Ratio) Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 75 APPENDIX A Table A1. Morphometric measurements of 26 species of doves Species Mass (g) Sternum Length (cm) Furcula Length (cm) Mourning Dove 116.8 ±14.0 (19) 48.1 ±1.6 (16) 21.7 ±1.0 (14) Inca Dove 43.8 * 34.4 ± 0.5 (2) 16.1 ± 0.6 (4) Common Ground Dove 30.1 * 33.1 ± 1.8 (3) 15.6 ± 0.3 (3) White-winged Dove 141.9 ± 10.3 (3) 51.6 ± 0.3 (2) 25.6 ±1.8 (2) Rock Dove 336.4 ± 29.3 (6) 62.7 ± 4.4 (4) 33.0 ± 0.4 (4) White-tipped dove 145.9 ±11.2 (3) 52.5 ± 2.1 (3) 25.1 ± 1.2 (2) Columba fasciata 399.0 ± 13.8 (6) 66.9 ± 2.4 (5) 36.1 ±1.5 (5) Columbina talpacoti 47.1 ± 0.6 (3) 35.5 ± 1.0 (3) 17.61 ± 0.3 (3) Caloenas nicobarica 492 * 7 0.9 37.2 Oena capensis 46.5 * 32.6 ± 1.9 (2) 16.1 ± 0.7 (2) Columba flavirostrls 250.9 ± 70.1 (3) 60.4 ± 1.1 (2) 30.5 ± 1.2 (3) Ocyphaps iophotes 205.5 ± 0.7 (2) 55.16 (1) 28.1 ± 2.1 (2) Columba subvinacea 180.0 51.0 2 6 .2 Columba guinae 352 66.6 35.9 Columba leucocephala 240.0 ± 12.7 (2) 54.8 ± 1.2 (2) 50.2 (1) Columba palumbus 538.0 ± 46.8 (3) 75.1 ± 3.3 (3) 40.8 ± 0.3 (3) Ptilinopus rivoli 141.0 39.0 Streptopelia orientalis 259.2 ± 15.9 (4) 60.4 ± 2.0 (4) 30.2 ± 0.7 (3) Streptopelia deaocto 146.0 56.8 2 7 .4 Ducula pacifica 386.4 ± 40.0 (5) 47.4 ± 1.8 (5) 36.3 ± 5.2 (4) Columbina minuta 36.4 ± 1.1 (2) 32.0 ± 0.4 (2) 15.2 ± 0.8 (2) Ducula bicolor *507.5 ± 34.6 (2) 63.9 ± 2.6 (2) 39.3 ± 1.0 (2) LeptotUa casslnli 145.2 ± 6.9 (3) 50.3 ± 0.3 (3) 26.6 ± 0.2 (3) Claravis pretiosa 69.4 ± 4.0 (2) 41.0 ± 1.0 (2) 19.9 ± 0.3 (2) Leptotila rufaxilla 157.5 ± 36.0 (2) 54.0 ± 1.9 (2) 26.2 ± 0.1 (2) Goura cristata 2000 * 117.9 ± 7.6 (3) 65.4 ± 4.7 (3) * Mass from Dunning (1993) Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 76 Table A1 continued Scapula Length (cm) Coracoid Length (cm) Humerus Length (cm) Radius Length (cm) 32.4 ± 1.8 (15) 25.5 ± 2.1 (19) 31.6 ± 0.9 (18) 32.8 ± .0 (19) 22.3 ± 1.8 (4) 18.4 ± 0.7 (4) 20.8 ± 0.6 (4) 22.2 ± 0.3 (3) 22.8 ± 0.6 (3) 18.1 ± 0.3 (2) 20.8 ± 0.5 (3) 22.0 ± 0.9 (3) 35.6 ± 0.7 (2) 27.8 ± 0.7 (2) 35.3 ± 1.5 (3) 37.2 ± 0.4 (3) 43.5 ± 2.5 (5) 36.6 ± 1.3 (5) 46.1 ± 3.1 (6) 49.9 ± 2.5 (6) 37.4 ± 1.5 (3) 28.3 ± 0.6 (3) 34.0 ± 1.5 (3) 34.4 ± 1.1 (3) 49.1 ± 1.0 (5) 39.4 ±1.1 (5) 47.9 ± 2.8 (6) 50.3 +1.5 (5) 25.5 ± 0.2 (3) 19.4 ± 0.1 (3) 22.6 ± 0.3 (3) 23.2 ± 0.1 (3) 55.7 ± 1.2 (2) 48.8 60.9 ± 0.2 (2) 66.6 ± 2.2 (2) 22.5 ± 0.8 (2) 17.8 ± 0.3 (2) 21.5 ± 0.2 (2) 24.1 ± 0.2 (2) 42.9 ± 0.3 (2) 33.8 ± 1.4 (3) 41.8 ± 2.1 (3) 44.1 ± 2.5 (3) 39.9 ± 1.0 (2) 30.3 ± 0.9 (2) 37.1 ± 0.9 (2) 35.8 ± 1.3 (2) 3 6 .0 2 8.5 3 6.6 37.9 43.1 3 6.9 4 8 .3 48.5 42.0 (1) 32.9 ± 1.8 (2) 41.9 ± 1.7 (2) 43.6 ± 1.8 (2) 51.9 ± 1.8 (2) 42.7 ± 0.3 (3) 54.9 ± 1.2 (3) 55.9 ± 1.0 (3) 3 5.6 29.0 33.0 3 5.5 41.5 ± 0.6 (4) 32.8 ± 0.8 (4) 41.7 ± 0.9 (4) 43.6 ±1.1 (4) 3 8.8 30.1 3 8.2 39.8 43.0 ± 2.4 (5) 39.0 ± 1.0 (4) 56.7 ± 1.7 (4) 60.2 ± 1.9 (4) 22.8 ± 0.7 (2) 17.5 ± 0.1 (2) 20.6 ± 01 (2) 20.9 ± 0.6 (2) 48.2 ± 3.2 (2) 43.5 ± 2.3 (2) 54.8 ± 1.7 (2) 58.8 ± 1.5 (2) 36.1 ± 0.1 (3) 27.7 ± 0.3 (2) 32.8 ± 0.4 (3) 34.3 ± 0.1 (3) 26.3 ± 3.9 (2) 22.7 ± 0.2 (2) 26.2 ± 0.6 (2) 28.7 ± 0.8 (2) 37.2 ± 0.1 (2) 29.1 ± 1.3 (2) 33.5 ± 0.4 (2) 35.2 ± 1.2 (2) 67.9 ± 33.6 (3) 76.3 ± 4.7 (3) 105.0 ± 4.3 (3) 117.2 ± 1.4 (3) Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 77 Table A1 continued Ulna Length (cm) Manus Length (cm) Pec Scar Area (cm^) 36.4 ± 1.1 (19) 22.9 ± 0.8 (19) 2.1 ± 0.3 (17) 25.3 ± 0.7 (3) 14.4 ± 0.2 (4) 0.9 ± 0.1 (3) 24.5 ± 0.7 (3) 14.5 ± 0.4 (3) 0.8 ± 0.0 (3) 42.6 (1) 25.6 ± 0.3 (3) 2.4 ± 0.4 (2) 55.1 ± 3.0 (5) 34.5 ± 1.6 (6) 3.8 ± 0.4 (4) 38.4 ± 1.3 (3) 23.6 ± 1.3 (3) 2.5 ± 0.4 (3) 56.5 ± 1.5 (5) 33.9 ± 1.0 (6) 3.4 ± 0.5 (4) 26.2 ± 0.0 (3) 14.8 ± 0.2 (3) 1.0 ± 0.2 (3) 74.7 ± 0.9 (2) 42.4 ± 0.5 (2) 6.2 26.8 ± 0.1 (2) 14.9 ± 0.8 (2) 1 .2 48.7 ± 2.0 (3) 29.2 ± 2.1 (3) 3.4 ± 0.3 (2) 39.7 ± 1.2 (2) 24.5 ± 1.3 (2) 2.6 ± 0.5 (2) 4 1.9 2 5.6 2.4 55.3 34.9 4.4 - 30.2 (1) 2.6 ± 0.2 (2) 62.9 ± 1.5 (3) 38.3 ± 0.5 (3) 5.1 ± 0.4 (3) 38.1 2 0.9 1.3 49.0 ± 0.8 (2) 29.8 ± 0.7 (4) 3.5 ± 0.2 (4) 4 4 .9 27.7 3.5 65.5 ± 2.5 (4) 35.8 ± 0.7 (5) 1.7 ± 0.3 (5) 23.4 ± 0.4 (2) 13.3 ± 0.2 (2) 0.8 ± 0.1 (2) 65.4 ± 1.8 (2) 37.1 ± 1.5 (2) 3.6 ± 1.0 (2) 37.7 ± 0.3 (3) 22.9 ± 0.3 (3) 2.6 ± 0.2 (3) 3 1 .7 18.2 1.6 ± 0.0 (2) 38.9 ± 0.6 (2) 23.6 ± 0.1 (2) 2.5 ± 0.0 (2) 128.3 ± 2.7 (3) 65.8 ± 3.1 (3) 14.8 ± 2.4 (3) Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 78 Table A2. Morphometric variables for the wings of 26 species of doves (mean ± SD, Species Mass (g) Wing Area (cm'^) Wing Length (cm) Common Ground Dove 35 56.4 ± 4.0 (3) 11.65 ±0.3 (3) Inca Dove 48.1 ±3.7 (22) 64.5 ± 3.9 (23) 12.2 ± 0.7 (23) White-tipped Dove 172.5 ± 12.2 (14) 156.8 ± 7.4 (14) 18.6 ±0.7 (14) Mourning Dove 105.2+13.0 (5) 127.8 ± 6.9 (5) 19.1 ± 0.4 (5) White-winged Dove 148.1 ± 10.4 (33) 160.8 ± 13.2 (34) 20.5 ± 1.0 (34) Rock dove 349.6 ±39.1 (28) 311.6 ±24.6 (28) 30.5 ± 1.2 (28) Columba leukocephala 247 * 240.8 ± 9.0 (2) 24.6 ± 0.3 (2) Columba palumbus 490* 378.1 ±33.7 (3) 32.4 ± 2.0 (3) Columba guinea 352 * 283.5 ± 7.8 (2) 28.6 ± 0.3 (2) Geotrygon violacea 97.8 * 210.0 ± 6.7 (2) 21.4 ± .04 (2) Columba subvinacae 172 * 180.9 21.4 Columba flavirostrls 302 ± 0.0 (3) 206.7 ± 37.0 (3) 24.0 ± 1.0 (3) Columba fasciata 392 * 298.5 ± 47.0 (2) 27.8 ± 2.8 (2) Streptopelia orientalis 215 * 265.3 ± 17.7 (6) 27.3 ± 1.5 (6) Ducula pacifica 386 * 418.4 ±27.1 (7) 31.2 ± 1.2 (7) Ocyphaps Iophotes 205 * 165.3 ± 20.2 (3) 21.1 ± 1.1 (3) Ptilinopus poryphraceus 105 * 124.6 ± 15.1 (6) 17.1 ± 1.2 (6) Ptilinopus rarotongensis - 126.6 ± 11.4 (4) 15.9 ± 0.7 (4) Columbina minuta 36.5 ±0.1 (2) 55.9 ± 6.6 (2) 10.6 ± 0.8 (2) Caloenas nicobarica 492 389.8 30.3 Claravis pretiosa 69.5 ± 3.5 (2) 89.5 ± 7.7 (2) 14.7 ± 1.1(2) Columbina talpacoti 47.5 ± 0.7 (2) 64.6 ± 1.8 (2) 11.4 ± 0.2 (2) Ducula bicolor 507.0 ±34.6 (2) 298.6 ± 48.4 (2) 28.5 ± 2.3 (2) Leptotila cassinii 145.3 ± 7.0 (3) 141.9 ± 2.2 (3) 17.0 ± 0.5 (3) Leptotila rufaxilla 183 156.3 17.7 Oena capensis 46 70.6 12.6 * Mass from Dunning (1993) $ Body width estimated from least-squares regression through the measured body widths from doves captured in the field. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 79 Table A2 continued Aspect Ratio Index Base Wing Chord (cm) Body Width (cm) Total Wing Area (cm) 4.8 ±0.1 (3) 6.3 ± 0.3 (3) 2.7 t 130.0 ±8.7 (3) 4.6 ± 0.4 (23) 6.7 ± 0.2 (23) 2.8 ± 0.3 (23) 147.6 ± 7.6 (23) 4.4 ±0.2 (14) 10.7 ± 0.6 (14) 4.8 ±0.3 (14) 365.7 ± 17.3 (14) 5.7 ± 0 .4 (5) 8.7 ± 0.3 (5) 3.6 ±0.1 (5) 286.7 ± 16.7 (5) 5.3 ± 0.4 (34) 9.9 ± 0.5 (34) 4.2 ± 0.4 (34) 363.4 ± 28.2 (34) 6.0 ± 0.3 (28) 12.0 ± 0.6 (28) 7.3 ±0.8 (28) 710.2 ± 57.2 (28) 5.0 ± 0.0 (2) 12.4 ± 0.5 (2) 3.9$ 530.1 ±20.0 (2) 5.5 ± 0.2 (3) 14.8 ±0.6 (3) 7.6 $ 868.6 ±71.8 (3) 5.8 ±0.1 (2) 12.9 ±0.1 (2) 5.5 t 638.0 ± 14.8 (2) 4.4 ± 0.2 (2) 11.6 ± 0.0 (2) 1.7 $ 439.4 ± 13.5 (2) 5.1 10.6 2.8 $ 391.4 5.6 ± 0.5 (3) 11.1 ± 1.2 (3) 4.8 $ 465.9 ± 77.9 (3) 5.2 ± 0.2 (2) 14.1 ±0.5 (2) 6.1 $ 682.9 ± 97.0 (2) 5.6 ± 0.4 (6) 12.1 ±0.5 (6) 3.4 $ 572.0 ± 35.5 (6) 4.7 ± 0.3 (7) 16.5 ±0.9 (7) 6.0 t 936.1 ±59.0 (7) 5.4 ± 0.3 (3) 9.9 ± 0.6 (3) 3.3 $ 363.1 ±42.3 (3) 4.7 ± 0.3 (6) 9.3 ± 0.2 (6) 1.8 t 265.7 ± 30.5 (6) 4.0 ± 0.2 (4) 9.7 ± 0.3 (4) -- 4.0 ± 0.2 (2) 6.7 ± 0.2 (2) 0.8 $ 116.8 ± 13.3 (2) 4.7 15.7 7.6 $ 899.1 4.8 ± 0.5 (2) 7.6 ± 0.3 (2) 1.3 t 188.5 ± 14.7 (2) 4.0 ± 0.0 (2) 7.2 ± 1.0 (2) 0.9 t 135.7 ± 2.5 (2) 5.4 ± 0.0 (2) 14.4 ± 0.6 (2) 7.8 $ 710.0 ± 108.8 (2) 4.1 ± 0 .2 (3) 10.1 ±0.3 (3) 2.4 $ 308.0 ± 3.3 (3) 4 10.9 3.0 $ 345 4.5 6.9 0.9 $ 147.4 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 80 Table A2 continued Wing Loading (Nm"^) Wing Disk Loading (Nm’O Wingspan (cm) Aspect Ratio 26.5 ± 1.7 (3) 6.4 1 0.3 (3) 26.1 1 0 .7 (3) 5.210.1 (3) 32.013.1 (22) 7.8 1 2.0 (23) 27.31 1.4 (24) 5.010.3 (23) 46.5 1 3.5 (14) 7 .9 1 1.2 (5) 42.1 1 1.4 (14) 4.910.2(14) 37.61 5.8(5) 8.8 1 0.3 (6) 41.8 1 0 .9 (5) 6.1 1 0 .4 (5) 40.2 1 2.6 (33) 12.21 1.2 (14) 45.3 12.1 (34) 5.7 1 0.4 (34) 48.4 1 5.0 (28) 9.41 1.1 (28) 68.2 1 2.7 (28) 6.6 1 0.3 (28) 45.7 1 1.7 (2) 10.9 1 0.3 (2) 53.1 1 0 .7 (2) 5.3 10.1 (2) 5 5 .6 1 4 .4 (3) 11.8 1 1.3 (3) 72.3 1 4.0 (3) 6.0 1 0.2 (3) 54.1 1 1.3 (2) 11.2 1 0 .2 (2) 62.7 1 0.6 (2) 6.2 1 0.0 (2) 2 1 .8 1 0 .7 (2) 6.2 1 0.0 (2) 44.4 10.1 (2) 4.5 1 0.2 (2) 43.1 10.3 45.6 5.3 64.9 1 12.4 (3) 13.61 1.5 (3) 52.8 1 1.7 (3) 6.0 1 0.6 (3) 56.918.1 (2) 13.012.3 (2) 61.6 1 5.5 (2) 5.6 1 0.2 (2) 37.0 1 2.3 (6) 8.0 1 0.8 (6) 58.0 1 2.9 (6) 5.9 1 0.4 (6) 40.6 1 2.4 (7) 10.4 1 0.7 (7) 68.3 1 2.3 (7) 5.0 1 0.3 (7) 55.916.1 (3) 12.41 1.1 (3) 45.6 1 2.2 (3) 5.7 1 0.3 (3) 39.2 1 4.9 (6) 10.3 1 1.6 (6) 35.9 1 2.4 (6) 4.9 1 0.3 (6) 30.8 1 2.9 (2) 9 .6 1 1.3 (2) 21.9 1 1.7 (2) 4.1 1 0 .2 (2) 53.7 13.2 68.3 5.2 36.3 1 4.7 (2) 9.4 1 2 .2 (2) 30.6 1 2.8 (2) 5.0 1 0.5 (2) 34.3 1 1.1 (2) 10.5 1 0.5 (2) 23.7 1 0.4 (2) 4 .2 1 0 .1 (2) 70.6 1 6.0 (2) 15.21 1.3 (2) 64.8 1 5.1 (2) 5.9 1 0.0 (2) 46.3 1 2.7 (3) 13.710.8 (3) 36.4 1 1.0 (3) 4.3 1 0.2 (3) 52 15.6 38.3 4.3 30.6 8.5 26 4.6 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 81 APPENDIX A FIGURE LEGENDS Figure A l. Whole-body kinematics of doves during slow flight (A. Common Ground Dove; B. Inca Dove; C. Mourning Dove; D. White-winged Dove; E. White-tipped Dove). Figure A2. Proposed phylogeny for the doves in this study. Phylogeny was adapted from Sibley and Alquist (1990; * branch length was documented to genus, $ slight difference was assumed between Common Ground Dove and Inca Dove). Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Figure Al.A. 82 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Figure Al.B. 83 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Figure Al.C. 84 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Figure AI D. 85 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Figure A I.E. 8 6 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Figure A2 87 ■Zenaida macroura 2.3 3.4 -Zenaida asiatica 5.4 -Leptotila verreauxi* ■Columba livia 8.2 . Columbina inca * 8.1 •Columbina passerina $ Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 88 APPENDIX B Flight Speeds Flight speeds of birds flying toward or away (± 10°) from me were recorded with a radar gun (Stalker Pro, .01s target acquisition time, 100m range). Locomotor Gaits in Doves Mourning Doves, White-winged Doves, and Rock Doves did not completely contract their wing during the upstroke in fast level flight (Figure ). Such a wingstroke cycle is assumed to be indicative of a bird using a continuous vortex gate in which lift is produced by the secondaries on the upstroke of the cycle (Tobalske and Dial 1996). Although the other species of doves may have used a similar gait, they were not recorded during fast level flight. Based on wake-vortex visualization, at least two gaits, the vortex-ring gait and the continuous-vortex gait, have been identified during flapping flight that are based on the aerodynamic function of the upstroke (Spedding et. al. 1984; Spedding 1987a, 1987b; Rayner 1991, 1995). In the vortex-ring gait, lift is produced only during downstroke, so a single vortex-ring is shed after each downstroke (Spedding et. al. 1984; Spedding 1987a, 1987b; Rayner 1991, 1995). In the continuous-vortex gait, lift is produced during both the downstroke and the upstroke, but the wings are slightly flexed during the upstroke, which gives rise to an undulating, continuous-vortex wake. Wingtip and wrist path during flight can be diagnostic of flapping gait (Scholey 1983; Rayner 1995; Tobalske and Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 89 Dial 1996). In the vortex-ring gait the wings are strongly flexed during upstroke; however, in the continuous-vortex gait the wrists and wingtips are not brought in close to the body, so lift may still be produced on upstroke. References Rayner, J.M. V. 1991. Wake structure and force generation in avian flapping flight. Contemporary Mathematics. 141; 351-400. Rayner, J.M.V. 1995. Dynamics of vortex wakes of swimming and flying vertebrates. In: C.P. Ellington and J. J. Pedley, eds. Symposia of the Society for Experimental Biology XLIX. Them Company of Biologists Limeted, Cambridge, 131-155. Scholey, K.D. 1983. Developments in vertebrate flight. Ph.D. thesis, Univ. Bristol, United Kingdom. Spedding, G. R. 1987a. The wake of a kestrel {Falco tinnunculus) in gliding flight. J. Exp.Biol. 127: 45-57. Spedding, O R. 1987b. The wake of a kestrel {Falco tinnunculus) in flapping flight. J.Exp.Biol. 127: 59-78. Spedding, G.R., Rayner, J.M.V , and C.J. Pennycuick. 1984. Momentum and energy in the wake of a pigeon {Columha livia) in slow flight. J.Exp.Biol. I l l : 81-102. Tobalske, B.W. and K.P. Dial. 1996. Flight kinematics of Black-billed Magpies and pigeons over a wide range of speeds. J.Exp.Biol. 199:263-280. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 90 Table B 1. Flight speeds of birds during straight and level flights (recorded with a radar gun; ms ') Species Flight Speed (Mean ± SD) Inca Dove 10.0 ± 1.3 (2) Mourning Dove 10.2 ± 1.9 (13) White-winged Dove 10.9 ± 1.3 (23) Common Ground Dove 7.4 ± 0.7 (5) White-tipped Dove 9.0 (1) Great-tailed Grackle (Male) 10.5 ± 1.4 (17) House Sparrow 9.3 ± 1.2 (3) Red-winged Blackbird 8.1 ± 1.9 (10) Laughing Gull 9 .4 Little Blue Heron 8 .2 Common Nighthawk 9.0 ± 2.2 (3) Scissor-tailed Flycatcher 7.5 ± 1.5 (3) Brown-headed Cowbird 1 0 .1 Northern Roughwing Swallow 9 .0 Dicksissel 14.3 ± 1.2 (7) Great Egret 8 .0 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 91 APPENDIX B FIGURE LEGENDS Figure Bl. Least-squares regression lines, RMA regressions, and correlation coefficients (r) describing the relationships between selected log transformed variables and log body mass for: Common Ground Dove; Inca Dove; Mourning Dove; White-winged Dove; White- tipped Dove; and Rock Dove. (A) Humerus length (cm); (B) radius length (cm); (C) scapula length (cm); (D) sternum length (cm); (E) coracoideus length (cm); (F) furcula length (cm); (G) wing length (cm); (H) aspect ratio index; (I) single wing area (cm^). Figure B2. Least-squares regression lines, RMA regressions, and correlation coefficients (r) describing the relationships between selected log transformed variables and log body mass for 26 species of doves. (A) Humerus length (cm); (B) manus length (cm); (C ) pectoralis scar area on the keel (cm“); (D) ulna length (cm); (B) scapula length (cm); (F) sternum length (cm); (G) furcula length (cm); (H) radius length (cm); (I) coracoideus length (cm); (J) wing length (cm); (K) total wing area (cm^); (L) single wing area (cm); (M) wing loading (Nm'“); (N) aspect ratio; (O) wing disk loading (Nm'^); (P) wingspan (cm), (note: for calculation of aspect ratio, wingspan, total wing area, wing loading and wing disk loading, body width was estimated from a regression line through body width measurements of 6 species of wild caught doves). Number guide to 26 species comparisons Mourning Dove 1 Inca Dove 2 White-winged Dove 3 Rock Dove 4 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 92 White-tipped Dove 5 Common Ground Dove 6 Columba fasciata 7 Columbina talpacoti 8 Caloenas nicobarica 9 Oena capensis 10 Columba flavirostris 11 Ocyphaps lophotes 12 Columba subvinacea 13 Columba guinae 14 Columba leucocephala 15 Columba palumbus 16 Ptilinopus rivoli 17 Streptopelia orientalis 18 Streptopelia deaocto 19 Ducula pacifica 20 Columbina minuta 21 Ducula bicolor 22 Leptotila cassinii 23 Claravis pretiosa 24 Leptotila rufaxilla 25 Goura cristata 26 Figure B3. Flight kinematics of doves during fast forward flight. (A) White-winged Dove lateral view wing kinematics; (B) Common Ground Dove lateral wing kinematics; (C ) Mourning Dove caudal view wing kinematics; (D) White-winged Dove caudal view wing kinematics. Figure B4. Flight kinematics of doves during slow forward flight. (A) White-winged Dove caudal view kinematics during post-takeoff flight; (B) White-winged dove lateral view kinematics; (C ) Mourning Dove lateral flight kinematics. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. FIGURE B l 93 50.1 y = 0J64x + 0.749 r * = 0.985 r = 0.992 Rock Dove 1.6 -39.8 1 ■B Wliitc-winged D ovc# J White-lipped Dove s 15- iVloiiiniiig Dove .31.6 30 I c 3 U 1.4 -25.1 Inca Dove Common Ground Dove lJ--« 20.0 13 2.0 (31.6) 1100.01 Log body mass (g) ( Body mass [g]) B y = 0363x + 0.772 r ^ = 0.981 r = 0.991 1.7 Rock Dove -50.1 1 -39.8 E W hiie-wuigeJ Dove # o t JS W) J • While-tipped Dove c Moiinii iig Dove J I ..5 -31.6 c2 I 1 1.4 25.1 Inca Dove ommon Ground Dove 1.3 20.0 1.5 2.0 2.5 (31.6) 11001 (316) Log body mass (g) (Body mass [g]) ReprocJucecJ with permission of the copyright owner. Further reproctuction prohibitect without permission. FIGURE B l. Continued. 94 so.i C y = OJlOx + 0.873 r ' = 0.961 r = 0.980 0 Rock Dove 1.6 ■ .39.8 Whitc-tipped Do\ e 0 White-wiiiged Dove %. I Mourning Dove i M 1.5 - -31.6 J Oùa U ca t(J 3 LO .25.1 S Inca Dove .ommon Ground Dove IJ . - 20.0 1.5 2.0 2.5 (31.6) ( 100 .0 ) (316.2) Log body mass (g) (Body mass (g]> D y = 0.293* + 1.073 r^ = 0.986 r = 0.993 1.8 -63.1 oS Whiie-tipped Dove Whiie-winged Duve* C -50.1 o .Mourning Dove 00c I I -39.8 # Inca Dove Common Ground Dove Ij H— 31.6 tJ 2.0 2j (31.6) (100-0) (316.2) Log body mass (g) (Body mass [g]) Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 95 FIGURE B 1 .Continued 1.6 39.8 y = OJiSx + 0.766 r 2 = 0.990 Rock Dove 1.S -31.6 While-winged Dove • Mourning Dove ■25.1 I bû J i 1.3 20.0 u Inca Dove Common Ground Dove 1.2 15.6 1.5 2.0 (31.6) 1100.01 Log body muss (g) (Body moss [g]) p l.6 •39.8 y = 0.334X-*■ 0.672 r ’= 0.982 r = 0.991 Rock Dove 1.5 - -31.6 I White-^MOged Dove While-tipped Dove -25.1 I Mourning Dove 20.0 # Inca Dove 1.2 15.8 Common Ground Dove 1.1 12.6 2.0 (31.6) ( 1 0 0 ,0 1 (316.2) Log body mass (g) (Body mass [g]) ReprocJucecJ with permission of the copyright owner. Further reproctuction prohibitect without permission. 96 FIGURE B 1. Continued. 31.6 Rock dove# y = 0.403% + 0.429 r z = 0.950 r = 0.975 .25.11.4 I ■£ Wlme-wingetl Do\e, - 20.0 Mourning Dove# #White-tipped Dove cW) U -15.8 60c - 12.6 ommon Ground Dove 1.0 . — 10.0 Ij 2.0 (31.6) (lOO.U) Log body mass (g) (Body mas.s [g]) 6.3 y = 0.074% + 0.558 rZ = 0.286 r = 0.534 Rock dove # Mourning Dove# 0.75 -5.6 Whiie-wingeJ Dove X T3 e 0.7 -5.0 Common Ground Dove a . i 0.65#Whire tipped Dove -4.5 0.6 -f- 4.0 1.5 2.0 (31.6) 100. 0 ) Log body mass (g) (Body m a.ss [g]) ReprocJucecJ with permission of the copyright owner. Further reproctuction prohibitect without permission. 97 FIGURE B 1. Continued. 2.6 ■398 y = 0.744% + 0.575 = 0.992 r = 0.996 Rock dove 2.4 - -251 es Whiie-wmgcU Dova # Wlme-iipped Dove -159 es 60 Mourning Dove 60 2.0 . Inca Dove .63.1 Common Ground Dove 1.6 39.8 13 23 (31.6) (316) Log body ma.ss (g) (Body ina.ss (gj) ReprocJucecJ with permission of the copyright owner. Further reproctuction prohibitect without permission. CD ■D Q.O C g Q. ■D CD C/) Log Humerus Length (cm) 3o ' Log Manus Length (cm) O 21 'ë 8 fS w w ci' 3 3" CD ■DCD C. O Q. aC O II 3 ■D O CD Q. <9- ■D CD C/) C/) Manus Length (cm) Humerus Length (cm) vO 00 CD ■D O Q. C g Q. ■D CD C/) Log Pectoralis Scar Area (cm ) o" d n 3 Log Ulna Length (cm) t 3? O s 'ë 3 dd 8 to ci' K p 3 (T> w Cl. 3. 3 " CD ■DCD .c- ? O Q. 3 aC ï O 3 ■D O CD Q. ■D CD C/) C/) Ulna Length (cm) Pectoralis Scar Area (cm ) VO CD ■ D O Q. C g Q. ■D CD C/) Log Scapula Length (cm) C/) Log Sternum Length (cm) 3 1 8 w 5 O to ci' O 3o 3 C D-CD 3. 3 " CD ■DCD O Q. C a OO 3O O "O « O 5: # CD Q. ■D CD C/) C/) Log Sternum Length (cm) Log Scapula Length (cm) 8 Figure B2. Continued. 101 10 100 y = 0.354X + 0.637 r2 = o.yyo r = «.yJ5 63.1 t 39.8 00 c ►J 1.4 25.1 i k U :• 10 15.8 10.0 2.(1 3.0 35 ( 10(11 (1000) (3162) L d» body mass (g) m ass [g]) 158 y = 0J97x + 0.700 r- = ().'((,S 26* 2.0 . . 100 uB OÜ • 16 c *7 J = 3 1.6 . 1 1.4 . .25 3.0 3.5 K ( 1000 ) (3162) 1.11% 1)11(1) lll.l\S (g) ' noivs |%|) Reproctucect with permission of the copyright owner. Further reproctuction prohibitect without permission. CD ■ D O Q. C g Q. ■D CD Log Coracoid Length (cm) Log Wing Length (cm) C/) 3o" O 8 ci' w CL 3 3 " CD ■DCD O c. CQ. a o 3 ■D O Q.CD ■D CD C/) C/) Log Coracoid Length (cm) Log Wing Length (cm) Figure B2. Continued. 103 3.0 1000 •2 0 y = 0.720s + 0.982 r-=i).9.M •2 2 -562 I i < -316 CbA .o O* 22 -17S o •21 1-100 :.u 3.0 iOO) (1000) i lu !>n(l\ iiiiss (g) üi'ih ni.i". i;;]) 562 L * y = 0.698% + 0.684 r- = o .o Uï 2.5 - - 316 C4 1 rt r4 2 2.2 - < • 12 1 00 C I00 1 I -63 13 3.0 <316) (1000) i.di' mass (g) HiKh ina.ss ([g]) ReprocJucecJ with permission of the copyright owner. Further reproctuction prohibitect without permission. Figure B2. Continued. 104 ■ 7 9 y = 0J178X + 1.014 r = = 0 . 6 f > 2 r = (t.S 14 C4 50 • 4 1.6 ■40 1 00c #6 1.4 .25 • IJ. _ 20 tJS 2.5 3.0 (31.6) (316) ( 1000) Lo^ liuil' mass (g) I Uiiih m ass [g]> ■63 y = 0.073*+ 0.531 r - = 0.’ftO r = 0.511) ,14 0.75- 5.6 O I 0.7 -s.o (S I K I 0.A5- -4.5 4.0 Z5 3.0 (316.2) (1000.0) Lou body mass (g) {liiidy mass [g]) ReprocJucecJ with permission of the copyright owner. Further reprocJuction prohibitecJ without permission. 105 Figure B2. Continued. -15.8 eu .15.6 .20.0 # 4 00 C 't 7.9 Û bû s: #6 .6.3 y = 0.194% + 0.585 r ' = 0.422 r = 0.650 0.7 Z5 .5.0 (316) (1000) Lui: Botls Miiss (g) (iiudy Mu.ss[g3) 1.9 y = 0.403% + 0.756 r - = 0.927 r = o.VW #22 63.1 1.7 -50,1 • 12 uE & #24 -31.6 1.4 -25.1 IJ 20.0 ZM (31,6) I HHh (lOOO) Lus l)iid> IIKISS (g) m a s s I g i ) Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Figure B3 106 Common Ground Dove White-winged Dove during fast forward during fast forward flig h t flig h t Mouring dove during fast forward flight White-winged Dove during fast forward flight (dorsal view) Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Figure B4 107 B White-winged Dove White-wing Dove during slow flight. during post-takeoff flight c Mourning Dove during slow forward flig h t Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 108 APPENDIX C Estimating Flight Muscle Mass Of Birds From Pectoralis Scar Area And The Muscle Mass Data Of Hartman (1961) Flight performance in birds is influenced by the mass of the flight muscles (Marden 1987, 1989, 1994; Ellington 1991; Seveyka 1999). For this reason, multiple species investigations of bird flight performance often include a direct measure of flight muscle mass (e.g., Marden 1987). However, it is not always practical to sacrifice the birds being studied in order to weigh the flight muscles, so an index of flight muscle mass must be used (Warrick 1998; Seveyka 1999). One index that has been used is the area of pectoralis scarring on the keel (Warrick 1998). Herein, I will present the comparison of pectoralis scar area to the flight muscle mass of Herons and Doves as measured by Hartman (1961). Scarring from the origin of the pectoralis on the keel was traced onto an acetate sheet from the skeletal remains of nine species of Herons and eight species of Doves. The traced areas were digitally scanned onto a computer and measured using NTH Image 1.6. Pectoralis scar areas were then compared to flight muscle masses calculated from Hartman’s (1961) flight muscle ratios for each species. Flight muscle mass was determined by multiplying the percent of body mass (flight muscle ratio) that consisted of the pectoralis, combined pectoralis and supracoracoideus, or total flight muscles (Hartman 1961) by the body masses of the Dove and Heron museum specimens. Based on linear regression models, pectoralis scar area appears to be a good index of flight muscle mass Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 109 for the two groups studied (the results of these comparisons are summarized in Tables D1 and D2 and Figures D 1 and D2). References Ellington, C.P. 1991. Limitations on animal flight performance. J. exp. Biol. 160:71-91. Hartman, 1961. Locomotor mechanisms of birds. Smithson, misc. Colins. 143: 1-91. Marden, J.H. 1987. Maximum lift production during takeoff in flying animals. J. Exp. Biol. 130: 235-258. Marden, J.H. 1990. Maximum load-lifting and induced power output of Harris’ hawks are general functions of flight muscle mass. J. exp. Biol. 149: 511-514. Marden, J.H. 1994. From damselflies to pterosaurs: How burst and sustainable flight performance scale with size. Am J. Physiol. 266: R1077-R1084. Warrick, D R. 1998. The turning-and liner-maneuvering performance of birds; the cost of efficiency for coursing insectivores. Can. J. Zool. 76: 1063-1079 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 110 Table Cl. Regression models comparing log pectoralis scar area (Pscar; cm^) on the keel to the log of pectoralis mass (Mp, combined pectoralis and supracoracoideus mass (M^J, and total flight muscle mass (M^^) calculated from Hartman (1961) for 7 species of herons and 9 species of doves. Log Muscle Group Least-squares model n r R F - Doves - Mp 1.056 + 1.329 (log Pscar) 8 0.990 0.980 297.9 Mps 1.163 + 1.256 (log Pscar) 9 0.989 0.978 312.7 1.228 + 1.365 (log Pscar) 8 0.987 0.974 226.8 - Herons - Mp 1.052 + 1.433 (log Pscar) 7 0.991 0.982 278.4 1.184-1- 0.953 (log Pscar) 8 0.994 0.988 413.7 1.237 + 1.500 (log Pscar) 6 0.975 0.951 76.8 Table C2. Mean Values used in regression models. Pectorals scar areas (Pscar) from skeletons and flight muscle ratios (FMR, %) from Hartman (1961). (FMRp = pectoralis FMR; M^ = pectoralis mass (g); FMR^^, = combined pectoralis and supracoideus FMR; M = combined pectoralis and supracoracoideus mass; Species Pscar (cm"^) Body Mass FMRp Mp FMRp, Mp, FMR,, Hh. (8) - Doves - Columba livia 3.8 300.0 20.3 60.9 23.5 70.5 30.9 92.7 Columba albilinea 3.4 310.0 21.3 66.0 25.2 78.0 34.6 107.3 Columba subvinacea 2.4 170.0 23.1 39.3 26.9 45.6 36.5 62.0 Columbina passerian 0.8 42.0 - - 29.3 12.3 - Columbina minuta 0.8 42.0 22.7 9.5 27.7 11.6 33.5 14.1 Columbina tlpacoti 1.0 46.5 22.8 10.6 28.1 13.1 34.1 15.8 Claravis pretiosa 1.6 69.0 26.2 18.1 30.9 21.3 39.0 26.9 Leptotila verreauxi 2.5 153.0 24.7 37.8 30.3 46.4 36.9 56.5 Leptotila cassini 2.6 148.0 28.0 41.4 34.2 50.6 42.4 62.7 H erons - Ardea herodias 9.4 1840.9 15.0 276.1 13.9 131.1 _ - Casmerodius albus 5.1 1041.6 13.0 135.8 13.6 69.6 22.3 232.2 Egretta thula 2.8 389.8 12.3 47.9 14.4 40.9 21.5 83.9 Egretta caerulea 3.1 370.3 14.5 53.7 14.8 46.4 22.3 82.5 Egretta tricolor 3.4 406.0 13.8 56.0 14.0 47.8 22.8 92.6 Butorides striatus 1.7 193.6 12.7 24.5 14.0 23.6 20.5 39.7 Bulbulcus ibis 2.3 287.5 14.9 42.7 16.4 38.3 23.5 67.7 Nycticorax nycticorax 2.8 417.9 - - 14.3 40.5 -- Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Ill APPENDIX C FIGURE LEGENDS Figure CL Least-squares regression line describing the relationship between log mean pectoralis scar area on the keel of doves (cm^) and the mean combined pectoralis and supracoraciodeus mass calculated from Hartman (1961). Figure C2. Least-squares regression line desribing the relationship between log mean pectoralis scar area on the keel of herons (cm^) and the mean combined pectoralis and supracoraciodeus mass calculated from Hartman (1961). Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 112 Figure Cl Log Combined Pectoralis and Supracoracoideus Mass (g) w p QO § w è - Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Figure C2 113 Log Combined Pectoralis and Supracoracoideus Mass (g) w w o s § K> ts orca i "j M' r O' bc s O Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 114 APPENDIX D WING KINEMATICS AND WHOLE-BODY ACCELERATION DURING FLIGHT IN ROCK DOVES {Columba livid) INTRODUCTION Slow speed wingbeat kinematics of birds can be classified into three major groups: symmetrical, asymmetrical (stroke and recovery), and wingtip-reversal (Norberg 1990). The major kinematic difference between the groups is the movement of the wing during upstroke. Hummingbirds use a symmetrical wingbeat in which the upstroke and downstroke are mirror images. The manus of a hummingbirds is strongly supinated and inverted during upstroke. Lift is produced during the upstroke and the downstroke of a symmetrical wingbeat, with the upstroke producing substantial lift (Norberg 1990). A stroke and recovery wingbeat cycle on the other hand is believed to only produce lift on the downstroke because wing is collapsed and pulled near the body during upstroke. A wingtip reversal upstroke is kinematically between the other two wingbeat cycles. The downstrokes in a wingtip reversal cycle is very much the same as that found in the other two types of wingbeats, however during upstroke the wingtip and wrist are kept away from the body and the hand wing is slightly supinated (Brown 1946). Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 115 Although the function of symmetrical and stroke-and-recovery wingbeats has generally been agreed upon, the role of the wingtip reversal is unclear. Brown (1946) first suggested that a wingtip reversal upstroke produces useful forces during flight in pigeons. No useful forces were reported for the start of upstroke and the dorsal extension of the wing, but the flick phase was touted as producing considerable lift and propulsion (Brown 1963). More recent evidence suggests that bats produce useful forces during upstroke (Aldridge 1986), and that a vortex ring is produced during the upstroke in Cockateils (Rayner and Thomas, unpublished results). In addition, Warrick and Dial (1998) found evidence of upstroke activity during slow flight maneuvering in pigeons. In order to gain a better understanding of the function of the wingtip reversal upstroke, we placed strain gauge accelerometers on Rock Doves to measure whole-body accelerations during the ascending flight. METHODS Three Rock Doves were captured in the wild using a remote-controlled trap. The doves were trained to fly to a perch for approximately 20 minutes a day for two months until they were capable of flying vertically to the top of a 2.5m perch while carrying two 2.5m long, six lead cables. To measure whole-body accelerations, two single axis strain gauge accelerometers (Entran; EGA-100-125) were imbedded perpendicular to one another in a 2.5cm x 1.5cm X 1cm piece of heavy neoprene from a computer wrist rest. An area of feathers in the center of the birds back were then trimmed so that they were < 1 cm in length. The Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 116 neoprene was then glued onto the back of the bird using self-catalyzing cyanoacrylate. The neoprene was positioned so that one accelerometer measured the acceleration along the axis of the body and the other measured acceleration perpendicular to the axis. Finally, two small (approximately Vi cm^) pieces of white tape were placed on the right side of the birds body, so that they formed a line parallel to the bird’s back. The accelerometers were zeroed individually before each trial by holding the bird in position so that the axis that the accelerometer measured was perpendicular to the axis of gravitational pull, thus, a single ‘g’ was registered when the accelerometers were rotated 90 degrees. Rock Doves were placed below the 2.5m perch and allowed to ascend and land at the top. During the flight, birds were video taped at 500Hz (Redlakes Motionscope 2000) with a level grid placed behind the bird. Signals from the accelerometers and syncronization signals from the video equipment were collect on an A/D system (National Instruments BNC-2090) at 5000Hz. Measurement were analyzed in Excel (Microsoft Inc.). Output voltages from the accelerometers were converted to accelerations values in g’s from the original calibration information for each accelerometer. From the video, body angle was determined by comparing the angle of the marks on the body to the level grid in the background. Body angles were used to determine the acceleration vector for each accelerometer and to translate accelerations with respect to the bird to accelerations with respect to horizontal. The resulting magnitudes and angles of acceleration from each accelerometer were added to determine the birds resultant acceleration vector with respect to horizontal. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 117 RESULTS The result from this experiment are summarized in Tables D1 and D2, and Figures D1 to D4. Discussion These data suggest that potentially useful forces are produced during upstroke in Rock Doves. Contrary to the results of Brown (1963), the accelerometer data presented here suggest that most of the potentially useful forces produced during the upstroke of Rock Doves in vertical flight are during early upstroke rather than during the flick phase. References Aldridge, H. D. J. N., 1986. Kinematics and aerodynamics of the greater horseshoe bat, Rhinolophus ferrumequimim, in horizontal flight at various flight speeds. J. exp. Biol. 126:479-497. Brown, R.H.J. 1948. The flapping cycle of the pigeon. J. Exp.Biol. 25:322-333. Brown, R.H.J. 1963. The flight of birds. Biol. Rev. 38:460-489. Norberg, U.M. 1990. Vertebrate flight: mechanics, physiology, morphology, ecology and evolution. Springer-Verlag, Berlin. Warrick, D R., and Dial, K.P. 1998. Kinematic, aerodynamic, and anatomical mechanisms in the slow maneuvering flight of pigeons. J. Exp. Biol. 201: 655- 672. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 118 Table D l. Angle and magnitude (R) of acceleration during wingbeat phase for three pigeons (mean ± SD; TU-BD = top of upstroke to bottom of downstroke; M D-BD = mid-downstroke to bottom of downstroke; BD-M U = bottom of downstroke to mid-upstroke; MU-TU = mid-upstroke to top of upstroke) P h ase Angle R Num ber of W ingbeats Eddy Summary (13 trials) TU-MD 101.1 ± 18.6 3.3 ± 0.3 25 MD-BD 12.6 ± 13.3 3.8 ± 0.3 24 BD-MU -6.2 ± 25.5 2.1 ± 0.4 29 MU-TU -39.6 ± 57.1 1.2 ± 0.4 26 Larson Summary (11 trials) TU-MD 97.6 ± 10.5 4.3 ± 0.3 19 MD-BD -1.1 ± 14.5 4.5 ± 0.8 21 BD-MU 8.7 ±29.2 1.8 ± 0.4 22 MU-TU -85.8 ± 21.0 1.6 ± 0.3 22 Niles Summary (12 trials) TU-MD 110.3 ± 9.1 4.5 ± 0.4 23 MD-BD -7.8 ±20.3 4.2 ± 0.5 23 BD-MU 23.9 ±26.2 2.2 ± 0.5 21 MU-TU -66.8 ± 27.9 1.6 ± 0.4 21 Table D2. Angle and magnitude (R) of acceleration during each wingbeat phase for combined data from three pigeons (mean ± SD; TU-BD = top of upstroke to bottom of downstroke; MD-BD = mid-downstroke to bottom of downstroke; BD-MU = bottom of downstroke to mid-upstroke; MU-TU = mid-upstroke to top of upstroke) Phase A n g le R # of Wingbeats SE of Angle SE of R TU-MD 103.1 ± 14.6 4 .0 ± 0 .6 65 4 .0 0 .0 8 MD-BD 1.0 ± 17.6 4.2 ± 0.6 67 4 .3 0 .0 7 BD-MU 6.9 ± 29.3 2.0 ± 0.4 71 5 .5 0 .0 5 MU-TU -61.8 ±44.8 1.5 ± 0.4 65 6 .8 0 .0 5 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 119 APPENDIX D FIGURE LEGENDS Figure D l. Approximate flight path and whole-body kinematics of a pigeon during vertical flight. Vectors indicate approximate, relative magnitude and angle of whole-body acceleration corresponding to the body position at that time. Figure also indicates the wing position used to categorize a wingbeat cycle into separate phases. Figure D2. Polar diagram indicating whole-body acceleration during separate phases of the wingbeat cycle. Position of bird is at top of upstroke. Dotted vector indicates the average flight path of the pigeon. Vectors illustrate the magnitude and direction of average acceleration during 5 phases of the wingbeat cycle: TU-MD = top of upstroke to mid- downstroke; MD-BD = mid-downstroke to bottom of downstroke; BD-preMU = bottom of downstroke to point just prior to mid-upstroke; preMU-preTU = just prior to mid-upstroke to just before top of upstroke; preTU-TU = phase from late upstroke to top of upstroke. Vector width represents the % of the wingbeat cycle averaged to produce each vector. Next to each mean vector is the angle (degrees) with respect to horizontal and magnitude (1 g = 9.81 ms'^) of the acceleration vector, and the percent of a wingbeat cycle represented by the vector. Figure D3. Polar diagrams illustrating the acceleration vectors (angle and magnitude) during a single wingbeat of (A) a representative sample (500Hz) indicating the number and position of samples in each phase; (B-D) acceleration profiles from three pigeons during vertical flight. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 120 Figure D4. Each polar diagram illustrated acceleration during four phases of a single wingbeat for four trials of vertical flight from three pigeons. Lines in the center of the polar plots represent the mean body angle during ascent. Arrows at the top of the graph indicate the mean angle of ascent. Note the similarity between trials in a single bird and between birds. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Figure D l. Top of Upstroke ^ ^ 1 Laie Upstroke Mid-upstroke Early Upstroke Bottom of Downstroke Mid-downstroke Early Downstroke Beginning of Downstroke Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Figure D2 122 ^ Flight Path 90 T U -M D 117 o 3.87 g 37% ot cycle preMU-preTU M D -B D 155 o 50 1.46 g 3.95 g 2 2 % o f cycle 15 % o f cycle 180 0 B D -preM U 349 0 preT U -T U 2.8 g 302 o 18.5 % of cycle 7% of cycle 270 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Figure D3. 123 O o il m • • o p* o o I? o o Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Figure D4. 124 NïlesROlwbl 9 0 90 120 60 120 60 g I m 30 S I 210 330 210 330 240 300 240 300 270 X bd-nrwr * lu-mdr # tu-mdr X bd-mur o ind-bdr 270 o md-bdr ▼ mu-tur 90 90 120 120 30 # # 210 330 210 330 240 300 240 300 • tu- met 270 » md- btt 270 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Figure D4. Continued. 125 90 UO 60 90 120 30 m e# • # • * 210 330 210 330 240 300 240 300 # t u - m d r 270 9 md-bdr 270 120 60 120 60 •• 30 V 210 330 210 330 240 300 240 300 • t u - m d * 270 270 ♦ m d - h d Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 126 Figure D4. Continued. EddyROlwt»! Eddy R03wb1 120 120 3 s I I • • ■ 210 210 330 330 240 300 2 4 0 300 # tu - m d 2 7 0 • m d- b d 270 M b d m u r * tu > m d w m i- m r ______• m d b d Eddy R04wbl eddy rlOwbt 90 120 60 120 I I I i Ii 210 330 210 330 240 300 2 4 0 300 w mu-tur 270 270 # tu-mdr Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. )t He le. iL—— I I ERR—— I RPL—— I ADD—— I XPO—— I LBL—— )CLC: s t a t : n în te re c ; 6 w w VU V:» K eplaced; 20000309 Used: 20000309 Pype : a E L v l: ■ S r c e : I Audn: Ctrl: Lang: III JLvl; m Form: C onf; 0 Biog: MRec: C try : III C ent : GPub; Fict: 0 Indx: 0 )esc: ■ I l l s : a F e s t: 0 D tS t: I Dates: 1999, )40 *c MTG )41 0 eng )92 *b )49 MTGA LOO 1 Seveyka, Jerred J. Î45 14 The effects of body size and morphology on the flight behavior and escape Light performance of birds / *c by Jerred J. Seveyka. 260 *c 1999. JOO ix, 180 leaves, bound : *b ill. ; *c 29 cm. 500 Typescript. 502 Thesis (M.S.)—University of Montana, 504 Includes bibliographical references. 710 2 University of Montana—Missoula. *b Division of Biological Sciences. *b rganismal Biology and Ecology Program. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 127 APPENDIX E Flight Kinematics of the Black-billed Magpie and Ringed Turtle-Dove Black-billed Magpies and Ringed Turtle-Doves are similar in size (approx. 150g), yet their flight kinematics vary considerably. The results presented here are similar to those previously documented by Tobalske and Dial (1996). Ringed Turtle-Doves use a wingtip reversal upstroke during take-ofif and slow flight, while Black-billed Magpies use a stroke and recovery wingbeat. During level flight, wingbeat kinematics of the dove indicate the use of a continuous vortex gait, while those of the magpie indicate the use of a ring-vortex gait (Tobalske and Dial 1996). The kinematics of the two species are summarized in figures El to E6. Ringed Turtle-Dove exhibited unusual pauses during the upstroke in slow flight (Figure E6). This behavior has also been observed in Inca Doves (Seveyka pers. obs.). References Tobalske, B.W. and K.P. Dial. 1996a. Flight kinematics of Black-billed Magpies and pigeons over a wide range of speeds. J. Exp. Biol. 199:263-280. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 128 APPENDIX E FIGURE LEGENDS Figure El. Lateral views of the path of the wingtip and wrist with respect to the body during the flight of a Ringed Turtle-Dove (top) and Black-billed Magpie (bottom) during (A) landing and (B ) forward flight. (From 60 Hz video). Figure E2. Lateral views of the body angle, and the path of the wingtip and wrist of a Ringed Turtle-Dove (top) and Black-billed Magpie during (A) landing, and (B) slow forward flight. (From 60 Hz video). Figure E3. Caudal view kinematics of the path of the wrist and wingtip of a Ringed Turtle- Dove (A. during forward flight; B. during takeoff) and Black-billed Magpie (C. during forward flight (From 60 Hz video). Figure E4. Lateral view kinematics of a Black-billed Magpie during takeoff (From 60 Hz video). Figure E5. Caudal view of a Ringed Turtle-Dove during slow forward flight with wingspan plotted for each frame (From 60 Hz video; note that the upstroke is extended longer than expected). Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Figure El 129 A Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Figure E l. Continued. 130 B Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Figure E2. 131 A Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Figure E2. Continued. 132 B Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Figure E3, 133 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Figure E4 134 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 135 Figure E5. 9 8 c -J------1- 4------1------1------1 3 4 5 6 7 8 9 Frame # 1 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 136 APPENDIX F Scaling of Wing and Skeleton Morphology in Herons and Kingfishers Seventeen morphological measurements were obtained from museum specimens for 11 species of Herons (Ardiadae) and species of kingfishers (Alcedinidae). The lengths of the sternum, furcula, coracoideus, scapula, humerus, radius, ulna, and manus were measured as in Tobalske (1996). In addition, pectoralis size was estimated by tracing the area of origin of the pectoralis from scarring on the keel; this estimate only provides an index of muscle size and does not represent the physiological cross-sectional area of the bipennate pectoralis muscle (Alexander 1983). The traced areas were then digitally scanned onto a computer and measured using NIH Image 1.6. External wing measurements were acquired from museum specimens of dried, spread wings by video taping the spread wings next to a line of known length. Video images of each wing were downloaded onto a computer (Macintosh Quadra 950 using Screenplay, Apple, Inc.) and measurements were obtained using Image 1.6 (National Institutes of Health). Morphological measurements obtained were: single-wing area (a), single-wing length (1), wing-root chord (c), and aspect-ratio index, calculated as E * a'\ similar to Tobalske (1996)]. Body masses were either recorded from museum tags associated with the dried skeletons and spread wings, or, when these data were not available, an average body mass reported for the species from Dunning (1993) was used. All variables were log-transformed and analyzed using least-squares linear Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 137 regression models to examine the relationships between body mass and morphology (SPSS 6.0). In addition, correlations coefficients and reduced-major-axis (RMA) regression coefficients were calculated to account for the presence of measurement error and statistical variation in both the independent and the dependent variables (Sokal and Rohlf 1981; Rayner 1985). RMA regression coefficients were calculated by dividing the least-squares linear regression coefficient (computed using Cricket Graph III 1.5, Comp. Assoc. Int., Inc., 1992), by the correlation coefficient (r). References Alexander, R. M. 1983. Animal Mechanics. Funtington, Chechester: Packard Publishing Ltd. Rayner, J.M.V. 1985. Linear relations in biomechanics; the statistics of scaling functions. J. Zool. Lond. 206:415-439. Sokal, R.R. and Rohlf, F.J. 1981. Biometry, 2nd ed. New York: W.H. Freeman and Co. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 00 c en g & Species Mass Sternum Length (cm) Furcula Length (cm) Scapula Length (cm) % Ardea herodius 1840.8 ±310.8 (7) 104.1 ±4.1 (7) 72.1 ± 1.9 (6) 98.9 ± 3.8 (7) 2 Casmewdius albus 1041.6 ±70.3 (5) 79.0 ± 2.6 (5) 53.1 ±2.5 (5) 73.1 ± 2 .0 (3) Q. 0C Egretta thula 389.8 ± 37.0 (5) 54.3 ± 3.3 (5) 37.2 ± 2.2 (3) 50.3 ±2.1 (5) ü3 Egretta caerulea 370.3 ± 7.3 (4) 56.4 ± 2.2 (4) 37.6 ± 1.4 (4) 50.8 ±2.1 (4) 1 (ÜQ. Egretta tricolor 406.0 ± 46.9 (5) 56.5 ± 3.2 (5) 39.5 ± 0.9 (5) 51.7 ± 1.5 (4) tr Butorides striatus 193.6 ± 16.8 (5) 40.9 ± 2.0 (5) 31.5 ±1.5 (4) 42.1 ± 0.7 (4) 3 Bulbulcus ibis 287.5 ±96.1 (4) 50.0 ± 1.0 (4) 37.4 ± 1.6 (4) 48.1 ± 1.4 (4) CCD O Nyctanassa violacea 689.3 ±46.2 (3) 64.2 ± 1.2 (2) 50.8 ±2.1 (2) 60.8 ± 1.9 (2) p Nycticorax nycticorax 808.2 ± 166.7 (5) 63.7 ± 2.8 (5) 54.2 1.6(5) 65.5 ± 1.9 (4) à 8 CD OC co & s 3 I û . Û:CD o> m CO CO CD Q. Table FI. Continued Coracoideus Length Humerus Length Radius Length Ulna Length (cm) Manus Length (cm) Pectoralis Scar Area "O (cm) (cm) (cm) (cnf) CD 78.1 ±2.3 (7) 194.6 ±7.7 (7) 216.6 ±6.8 (7) 226.0 ± 7.9 (7) 101.5 ±2.7 (7) 9.4 ± 1.8 (7) 2 58.8 ± 2.6 (5) 149.4 ± 5.2 (5) 169.4 ±5.1 (5) 175.3 ± 5 .6 (4) 79.2 ± 3.0 (5) 5.1 ± 0.5 (5) Q. 39.6 ± 1.4 (5) 91.2 ± 3 .4 (5) 105.0 ± 3.5 (5) 109.3 ± 3.3 (4) 52.0 ±2.1 (4) 2.8 ± 0.3 (5) O "O3 2 40.6 ± 0.7 (4) 93.4 ± 0.2 (4) 104.9 ± 0.6 (4) 109.6 ± 0.8 (4) 53.0 ± 1.4 (4) 3.1 ± 0.3 (4) Q. CD 41.1 ± 1.6 (5) 95.6 ± 4.6 (5) 107.2 ± 4.2 (4) 111.0 ± 5.6 (3) 55.8 ± 1.9 (3) 3.42 ± 0.5 (4) CD ■c 35.3 ± 2.3 (5) 67.2 ± 3.9 (5) 71.2 ± 3.6 (5) 74.6 ± 3.8 (5) 37.0 ±2.2 (5) 1.7 ± 0.5 (4) 3 39.7+1.1 (4) 88.8 ± 2.8 (4) 98.4 ± 2.0 (4) 102.5 ± 1.9 (4) 46.3 ± 1.3 (4) 2.3 ± 0.2 (4) 52.7 ± 0.7 (2) 109.4 ± 4.6 (2) 120.7 ±4.5 (2) 126.4 ±5.1 (2) 60.9 ± 3.3 (2) 3.8 ± 0.3 (3) ■P 55.5 ± 2.3 (5) 117.9 ±4.7 (5) 121.9 ± 4.4 (5) 128.2 ± 5.2 (5) 64.2 ± 3.2 (5) 2,8 ± 0.4 (5) 8 CO CO CD Q. "O 8 "O3 2 Q. CD q : ? (/>(/) CD Q. Table F2. Morphometric variables for the wings of nine species of Herons (mean ± SD) Species n Mass Wing Area (cm^) Wing Length Aspect Ratio Base Wing Chord "O (cm) Index CD 4 2390 1589.0 ± 228.4 67.9 ±3.1 2.9 ± 0.2 27.8 ± 3.6 2 Ardea herodias Q. C Bulbulcus ibis 5 338 476.3 ± 18.7 36.7 ± 1.2 2.8 ±0.1 15.8 ± 0.6 o "G 6 212 262.6 ± 28.9 26.6 ± 1.7 2.7 ±0.1 12.2 ± 1.1 "O3 Butorides striatus 2 Q. Casmerodius albus 3 874 1221.4 ± 138.2 58.4 ± 3.3 2.8 ±0.1 25.3 ± 1.6 2 Egretta caerulae 5 339 568.0 ± 43.3 40.9 ± 1.7 2.9 ± 0.1 17.4 ±0.7 ■c Egretta sacra 6 356 665.0 ± 43.3 41.7 ±2.0 2.6 ± 0.2 19.3 ± 1.5 Egretta tricolor 5 374 501.5 ±49.6 38.4 ± 1.6 3.0 ±0.1 15.6 ± 1.1 Nyctanassa violaceus 3 683 661.1 ±61.4 42.5 ± 1.6 2.7 ±0.1 19.7 ± 1.4 g Nycticorax nycticorax7 883 835.0 ± 47.9 47.8 ± 2.5 2.7 ±0.1 21.8 ±0.9 8 O g '(/)(/> CD Q. "O 83 "O 2 Q. CD Q1 141 Table F3. Morpbometric variables for bones of 10 species of kingfishers (mean ± SD, with number of birds in parentheses). Species Mass (g) Sternum Length Furcula Length (cm) (cm) Belted Kingfisher 156.5 ± 14.0 (13) 40.2 ± 2.3 (12) 23.5 ± 1.2 (12) Ringed Kingfisher 317 * 48.7 ± 1.0 (3) 30.3 ± 1.1 (3) Green Kingfisher 37.5* 22.1 ±0.8 (7) - Pygmy Kingfisher 14.7 ± 1.15 (3) 16.1 ±0.7 (3) - Halcyon sancta 44.0 19.7 16.5 Alcedo atthis 27.7 20.8 15.4 Halcyon chloris 31.6 ± 3.0 (3) 19.4 ± 0.4 (3) 18.5 ± 0.2 (3) Halcyon megarhyncha55.6 ± 10.4 (2) 17.1 ±0.7 (2) 19-4 Halcyon ruficollaris 53.7 ± 11.1 (3) 16.1 ±0.5 15.8 Halcyon tuta 42.3 ±4.1 (4) 16.4 ± 0.6 (4) - Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 142 Table F3. Continued. Scapula Length Coracoid Length Humerus Length Radius Length (cm) (cm) (cm) (cm) 37.5 ± 1.6 (12) 30.4 ± 2.3 (13) 48.8 ± 1.5 (12) 56.4+ 1.7 (12) 44.8 ± 0.6 (3) 40.1 ± 0.5 (2) 61.3 ± 1.3 (3) 72.5 ± 1.5 (3) 22.2 ± 0.5 (6) 19.8 ±0.5 (7) 26.7 ± 1.1 (7) 30.4 ± 0.9 (7) 16.7 ± 0.3 (3) 15.3 ± 1.0 (3) 19.1 ±0.4 (3) 22.1 ± 0.3 (3) 22.8 21.7 30.5 37.2 20.1 18.9 23.0 - 24.4 ± 0.3 (3) 24.7 ± 0.2 (3) 33.4 ± 0.7 (3) 40.0 ± 0.8 (3) 22.9 ± 0.1 (2) 22.5 ± 0.2 (2) 27.8 ± 0.0 (2) 32.8 ± 1.1 (2) 22.1 ± 1.0 (3) 20.3 ± 0.6 (6) 29.6 ± 0.4 (3) 34.9 ± 0.3 (3) 20.7 ± 0.8 (4) 19.3 + 0.9 (4) 27.9 ±3.1 (4) 33.3 ± 1.3 (3) Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 143 Table F3. Continued. Ulna Length Manus Length Pec. Scar Area (cm) (cm) (cm^) 60.0 ± 1.7 (12) 27.3 ±0.9 (13) 1.5 ±0.1 (12) 76.9 ± 0.2 (3) 33.9 ± 1.0 (3) 2.2 ± 0.3 (3) 32.6 ± 0.7 (7) 13.6 ± 0.4 (7) 0.44 ± 0.04 (6) 23.7 ± 0.3 (3) 9.3 ± 0.5 (2) 0.2 ± 0.00 (3) 39.1 15-5 0.43 29.5 13.3 0.36 42.3 ± 0.4 (3) 16.8 ± 0.8 (2) 0.43 ± 0.57 (3) 35.2 ±0.4(2) 14.3 ± 1.1 (2) 0.41 ±0.05 (2) 37.1 ±0.4 (3) 15.3 ± 0.0 (3) 0.27 ± 0.04 (3) 36.1 ±0.9 (3) 14.7 ± 0.4 (3) 0.26 ± 0.02 (4) Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 144 Table F4. Morphometric variables for the wings of 6 species of Species Mass (g)* n Wing Area (cm^) Alcedo atthis 27.7 5 43.9 ± 3.6 Ceryle aclyon 156.5 ± 14.0 ( 13) 4 175.7 ± 8.3 Chloroceryle atnericana 3 7 ^ + 2 54.2 ± 0.2 Halcyon chloris 31.6 ± 3.0 \3) 6 94.7 + 9.5 Halcyon ruficolaris 53.7 ± 11.1 (3) 4 89.5 ± 8.5 Halcyon tuta 42.3 ± 4.1 (4) 7 86.3 ± 4.6 * masses from skeletal specimens, ± Irom Dunning 1989. Table F4 continued. Wing Length Aspect Ratio Base Wing (cm) Index Chord 10.7 ± 0.6 5.3 ± 0.2 4.9 ± 0.2 22.8 ± 0.7 5.9 ± 0.2 9.9 ± 0.4 11.9 ± 0.5 5.3 ± 0.4 4.9 ± 0.1 15.1 ± 1.2 4.8 ± 0.4 7.6 ± 0.4 13.9 ± 0.8 4.3 ± 0.3 8.0 ± 0.5 13.2 ± 0.6 4.0 ± 0.2 7.9 ± 0.3 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 145 APPENDIX F FIGURE LEGENDS Figure FI. Least-squares regression lines and regression (r^) and correlation coefficients (r) describing the relationships between selected log transformed skeleton variables and log body mass for eleven species of herons (lengths in cm). (A) furcula length (cm); (B) pectoralis scar area (cm^); (C ) radius length; (D) scapula length; (E) humerus length; (F) manus length; (G) sternum length; (H) ulna length; (I); coracoideus length (cm). Figure F2. Least-squares regression lines and coefficients (r^) describing the relationships between selected log transformed wing variables and log body mass for eleven species of herons. (A) wing length (cm); (B) single wing area (cm^); (C ) aspect ratio index. Number guide to heron scaling figures Species Number Ardea herodius 1 Casmerodius albus 2 Ardea cinerea 3 Egretta thula 4 Egretta caerulea 5 Egretta sacra 6 Egretta tricolor 1 Butorides striatus 8 Bulbulcus ibis 9 Nyctacorax violacea 10 Nycticorax 11 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 146 Figure F3. Least-squares regression lines, regression (r^) and correlation coefficients (r) describing the relationships between selected log transformed skeletal variables and log body mass for 10 species of kingfishers. (A) coracoideus length (cm); (B) furcula length (cm); (C ) humerus length (cm); (D) pectoralis scar area on the keel (cm^); (E) manus length (cm); (F) radius length (cm); (G) scapulalength (cm); (H) sternum length (cm); (I) ulna length (cm). Figure F4. Least-squares regression lines and coefficients (r^) describing the relationships between selected log transformed wing variables and log body mass for six species of kingfishers. (A) single wing area (cm“); (B) wing length (cm); (C ) aspect ratio index. Number guide to kingfisher scaling figures Species Number Belted Kingfisher (Ceryle alcyon) 9 Ringed Kingfisher ( Ceryle torquata) 10 Green Kingfisher (Chloroceryle americana) 3 Pygmy Kingfisher (Chloroceryle aenea) 1 Halcyon sancta 5 Alcedo atthis 2 Halcyon chloris 8 Halcyon megarhyncha 7 Halcyon ruficollaris 6 Halcyon tuta 4 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Figure FI. 147 1.9 79.' A y = 0.3S5X + 0.678 = 0.963 r = 0.981 1.8 63.1 S o CJ 1.7 S0.1 bû C 3 J -2 39.8 ü 1.5 31.6 1.4 25.1 2.1 2.4 2. 3.0 3.3 (125.9) (251.2) (5(fl.2) (1000) (1995) Log Body Mass (g) (Body Mass[g]) 1.0 100 B y = 0.617% - 1.160 r ^ = 0.848 r = 0.921 E 0.8 .6.31 a n uS I Kl 5 1 N c/i 0.6 3.98 15 2 So (X hJ 0.4 -2.51 0.2 1.58 2.10 2.40 2.703.00 3.30 (125.9) ( 2 5 1 2 ) (501.2) ( 1000 ) ( 1 9 9 5 ) Log Body M ass (g) (Body Mass [g]) ReprocJucecJ with permission of the copyright owner. Further reproctuction prohibitect without permission. 148 Figure FI. Continued. 2 . 4 2 5 1 C y = 0.450% + 0.849 r : = 0.947 r = 0.973 2.3 -200 S o -158 •S bO .126 c J en '•S.2 .100 « û5 1.9 -79.4 63.1 2.10 2.40 2.70 3,00 3.30 (125.9) (251.2) (501.2) (1000) (1995) Log Body M ass (g) (Body M ass [g]) 2.0 too D y = 0353% + 0.801 = 0.969 r = 0.984 1.9 79.4 E U U cW) J 1.8 -63.1 a 3 ra u o« (/! 00 ûû '• 7 1.7 -50.1 1.6 39.8 2.1 2,7 3.0 3.32.4 (125.9) (251.2) (501.2) (1000) ( 1 9 9 5 ) Log Body Mass (g) (Body Mass [gj) Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Figure FI. Continued. 149 2J 199.S E = 0.437X + 0.839 = 0.968 r = 0.984 158.S E U Ê 3 125.9 % W3 c g U (U in I 2.0, 100.0 2 (U X E ÛÛ 3 • 4 J X 1.9 - 79.4 1.8 63.1 2.1 2.4 2.7 3.0 3.3 (125.9) (251.2) (501.2) (1000) (1995) Log Body Mass (g) (Body Mass [g]) 2.1 126 F y = 0.415% + 0.639 r : = 0.973 r = 0.986 2.0 -100 1.9 E u£ CJ s 1.8 -63.1 bû OÛ 0>C J J on C /3 3 .50.1 c3c 1.7 § S 00 s 1.6 .J9 .8 1.5 .31.6 2.10 2.40 2.703.00 3.30 (125.9) (251.2) (501.2) (1000) ( 1 9 9 5 ) Log Body M ass (g) (Body Mass [g]) Repro(juce(j with permission of the copyright owner. Further reproctuction prohibitect without permission. 150 Figure FI. Continued. G 126 y = 0365x + 0.791 r * = 0.960 r = 0.990 2.0 -100 E CJ • 3 CJ bb 1.9 c -79.4 t U E 5 3E S B 1.8 I 1.7 1.6 J9.8 2.1Ô 2,70 3,00 3.30 (125.9) (501.2) (1000) (1995) Log Body Mass (g) (Body Mass [g]) 2.4 251 y = 0.441% + 0.990 r - = 0.950 r = 0.975 2.3 .200 E CJ 2.2 .158158 o ■S ao bû c C J(U 2.1 10# J cd 1.9 .79.4 a# 1.8 63.1 2.10 2.40 2.7b 3.(j0 3.30 (125.9) (251.2) (501.2) (1000) (1995) Log Body Mass (g) (Body M ass [gJ) ReprocJucecJ with permission of the copyright owner. Further reproctuction prohibitect without permission. Figure FI. Continued. 151 1.9 r7 9 .4 I y = OJSlx + 0.720 r * = 0.967 r = 0.984 • 3 1.8 . -63.1 E u •2 uE W) -50.1 J an 143 ■o 'o u 1.6 - -39.8 I j --- 31.6 2.10 2.40 2.70 3.00 3.30 (125.9) (251.2) (501.2) (1000) (1995) Log Body Mass (g) (Body Mass [g]> ReprocJucecJ with permission of the copyright owner. Further reproctuction prohibitect without permission. Figure F2. 152 1.9 -79 A y = 0376X + 0.607 r ^ = 0.928 r = 0.963 1.8 .6 3 g E 00 u J 00c #10 ►J ÛÛ 1.6 .4 0 c: u • 9 1.4 + 2.1 2.4 2.7 3.0 3.3 (125.9) (251.2) (501.2) (1000) (1995) Log Body Mass (g) (Body Mass [g]) 3.2 -1585 B y = 0.743% + 0.790 r 2 = 0.942 • - CN 3.0 -1000 ÇJ£ O'! I üS cm S 2.8 • S -631 aû # 7 a, # 9 & 2.5 -316 2.2 159 2.1 2.4 3.02.7 3.3 (125.9) (251.2) (501.2) (1000) (1995) Log body mass (g) (Body mass [g]) Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Figure F2. Continued. 153 0.4g ■3.02 y = O.OlOx + 0.421 r 2 - 0.026 - 2.88 -2 .7 5 - a 10 1, u C. 2 6 3 2.51 2.7 3 3.3 125.9 (251.2 (S01.2) ( 1000) (1995) Log body mass (g) (Body mass [g]) Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Figure F3. 154 y = 0Jllx + 0.811 r^ = 0.965 r = 0.982 -39.S E -2S.1 c t/ï3 *2 ‘o 1.2 . • 15.8 Ë o U 1.0 10.0 1.0 1,5 2.5 3.0 (10.0) (31.6) (100.0) (316.2) (1000) Log Body M ass (g) (Body M ass [g]) 1.6 39.8 y = 0.285% + 0.756 = 0.937 B r = 0.968 s o 1.4 25.1 E so CJ J bû c ►J a U. 3 bû E O 1.2 15.8 U & 1.0 10.0 1.0 1.5 2.0 2.5 ( 10.0 ) (31.6) (100.0) (316.2) (l(w8) Log Body M ass (g) (Body M ass [g]) Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Figure F3. Continued. 155 1.8- y = 0,392% + 0.809 r ^ = 0.977 63.1 r = 0.988 39.8 j= S 'J 00c nJ 00 C/J c S • 6 u • 7 <0B 3E X 1.2 15.8 1.5 2.0 ( 10.0 ) (31.6) (100.0) (316.!) Log Body Mass (g) (Body M ass [g]) y = 0.825% - 1.761 r ^ = 0.848 r = 0.92 0.4 . .2 .5 1 CN - 1.0 V) 1 DU -0 .4 - 0.40 - 0.8 0.16 1.0 1.5 2.0 2.5 3.0 (10.0) (31.6) ( 100.0) (316.2) (1000) Log Body Mass (g) (Body Mass [g]) Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Figure F3. Continued. 156 1.6 3 9 . 8 y = 0.424X + 0.474 = 0.970 r = 0.985 1 .4 . -2 5 .1 E OE o cao J 1.2 15.8 U cai 1.0 - 10.0 0.8 4 - 6.3 1.0 1.5 2.0 2.5 ( 10.0 ) (31.6) ( 100.0) (316.2) Log Body Mass (g) (Body M ass [g]) 2.0 100 F y = 0.390% + 0.885 = 0.970 1.8 -63,1 I -Ju g 1.6 39.8 Cbû u -% Qd 1.4 25.1 1.2 -}- 15.8 1.0 1.5 2.0 2.5 (10.0) (31.6) ( 100.0 ) (316.2) Log Body Mass (g) (Body M ass [g]) Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Figure F3. Continued. 157 y = 0.337% + 0.800 r ' = 0.958 r = 0.979 1.6- 39.8 I rS50 C u 0 . 1 .4 - • 8 - 25.1 & • 7 1.2 4 - 15.8 1,0 2.0 2,5 3.0 (10.0) (31.6) (100.0) (316,2) ( 1000) Log Body Mass (g) (Body Mass [g]) 1.8 63.1 y = 0.387% + 0.664 r : = 0.711 H r = 0,843 • 10 1.6 - .3 9 .8 C 50 E C CJ —Iu 5 0 C -25-1 J 3E c/7B # 7 1.2 15.8 1.0 1.5 20 2.5 3,0 (10.0 ) (31.6) ( 100.0) (316.2) ( 1000 ) Log Body Mass (g) (Body Mass [g]) Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Figure F3. Continued. 158 2.0 100 I y = 0J88x + 0.916 r ' = 0.975 r = 0.987 .6 3 .1 u£ EU “ 1.6. -3 9 .S Oi) Jj c: ea c U 5 S 1.4. .25.1 1.24- 1S.8 1.0 1.5 2.0 2.5 3.0 (10.0) (31.6) (100.0 ) (316.2) (1000) Log Body M ass (g) (Body Mass [g]) Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Figure F4 2 .4 159 y = 0.657% + 0.812 r ^ = 0.722 1 : . 2 < ^ 2.0 g JU £ 1.8 C% a 1.6 2 • 1.2 1.5 1.8 2.0 2.2 1.4 y = 0.377% + 0.519 r : = 0.808 E B u 00 c J oo 1.2 c g 50 J 1.0 1.2 1.5 1.8 2.0 2.2 y = 0.090% + 0.538 r * = 0.151 O « OC 0.6 1.2 2.2 Log Body Mass (g) {Body Mass [gj) Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 160 APPENDIX G Learning Through Slow-Motion: Using Motion Analysis As A Tool For Exploration And Inquiry Jerred Seveyka, Cassie A. Shigeoka, and Ryan W. Bavis Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 161 Background From pets and pigeons to people, animals move around us every day. But many animal motions are too fast for us to fully appreciate. A useful method for bringing technology and inquiry into the laboratory (as recommended by BSCS 1993, NRG 1997) is the use of video, as well as motion and still cameras, to slow motions for easy and repeated viewing, and for motion analysis, or kinematics. Kinematic studies have been an important tool in understanding the relationship between an animal’s design and its motion, and the way in which different animals move in order to achieve the same goal (e.g., feeding or fleeing). A kinematic study may consist of collecting footage of an animal in motion, slowing the motion, and then following the movements of the animal or its body parts to quantify factors such as velocity, frequency of oscillation, joint angle, or foot placement. Such studies provide excellent opportunities for students to sharpen their observation skills, develop questions, and test hypotheses that relate animal motions to the design, size, age, behavior and ecology of an animal. With the appropriate equipment, motion analysis can be applied to subjects ranging from the amoeba to the athlete. Furthermore, kinematics may provide one alternative to dissection by allowing students to explore animal design through the noninvasive study of animal locomotion (e.g., Balcolmbe 1997). Here we discuss procedures for exploring motion analysis and offer suggestions for doing field and laboratory experiments using video. In addition, we discuss some of the information that can be acquired from studying a moving subject, give examples of kinematics, and provide suggestions for further areas of study. Although our discussion will be restricted to video analysis in kinematic studies, most of the techniques described here can be applied to motion or still camera (e.g., stroboscopic) photography. We discuss video because: 1) the equipment and the tapes for data collection are cost effective, 2) video cameras and VCRs are common in both homes Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 162 and schools, 3) video footage is easy to collect (the equipment is portable and it is easier to achieve correct exposure with video than with film), and 4) video does not require processing as film does. Furthermore, most video can be viewed at 30 frames/s, which is equal to or faster than can be captured inexpensively by motion picture or still photography. Collection of Video Data E q u ip m en t Video equipment requirements will vary depending on the subject being recorded. Most standard video is viewed at 30 frames/s, but some equipment will allow frames to be split and viewed at an equivalent of 60 frames/s (e.g., Panasonic AG-1960, SVHS player). The speed of the motion and the sampling rate of the video will determine the type of information that can be gained and how that information is interpreted. For example, when studying a fast motion like pecking in woodpeckers, a sampling rate of 30 frames/s may only document the beginning and end of a peck and give the illusion that a peck follows a straight line. By doubling the sampling rate (60 frames/s) the observer may notice that the peck actually follows an arc. Finally, increasing the sampling rate further with high speed video or film (125 to 10,000 frames/s) may provide detailed information on the full nature of the motion (Figure 1). Fortunately, many animal motions are slower than woodpecker pecking, and standard (30 frames/s) video can be applied to countless studies of motion. As in 35 mm photography, a high shutter speed must be used in order to stop motion clearly. If the subject is moving extremely fast (e.g., the wing of a sparrow in flight or the tip of a golf club during a swing) then a shutter speed of 1/2000 s or faster should be used to prevent blurring of the image (Note: A fast shutter speed requires more light than a slower shutter speed to produce images with the proper exposure). For most human motions, a shutter speed of 1/500 s or 1/1000 s is fast enough to produce a clear image. For many applications standard video formats will produce images that provide Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 163 adequate detail for motion analysis. However, if fine resolution measurements are needed or the subject represents a small portion of the entire image, a Hi-8 or Super-VHS (SVHS) recorder and the corresponding video tape will provide an image with greater resolution. Finally, if extremely clear images are needed, the best approach is to use a motion picture camera with a large film format (16 mm or 35 mm). Subjects for study Involving students in data collection (e.g., using them as the subjects or having them collect the video images) can be an easy way to increase student interest in kinematic projects. Subjects also can be found in pet stores, wildlife parks, zoos and aquariums, on campuses or playgrounds, and at sporting events. If a video recorder is not available, motion analysis can still be performed. Nature videos are full of high quality footage of animals in motion and often some of the motions are already slowed. The internet and CD-ROMS are also useful for finding video-clips of animals in motion. However, video from these sources may not include the video sampling rate, or may have been collected with a slow shutter speed; therefore the images may be blurred in freeze frame. In many cases, such as videotaping animals in the wild, the subjects will not be under control or in a controlled setting, which may result in poor angles of view that make documenting body movements difficult. Thus, the use of pets or people that can follow instructions and can be adorned with markers may provide more detailed information. Designing an experiment If you wish to understand a motion by plotting movements as they occur in an XY plane, then collect video footage with a stationary camera and a subject that only moves across the camera’s field of view. We will limit our discussion to analyzing video that was collected with such an experimental design*. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 164 Before beginning, there are several procedures that can improve the quality of the kinematic information gathered from video. Placing a grid behind a subject can provide a frame of reference with which measurements such as speed (distance moved over time) or acceleration (change in speed over time) can be made (Figure 2). For instance, if a subject walks in front of a grid of known dimensions, the boxes can be used to measure the distance the subject moved. If the time elapsed to cover this distance is also known (e.g., 3 frames of video, with a sampling rate of 30 frames/s, would be equivalent to 1/10 s), then the subject’s speed can be calculated. To increase the accuracy of your measurements from such a design, keep the camera as far away from the subject as possible, while keeping the subject as close to the grid as possible. If the subject is much closer to the camera than to the grid, the subject will appear to have moved farther (across more of the grid) than it actually has. An alternative to placing a grid in the background is to use a scale directly on the subject (for greater measurement accuracy this scale should be as large as possible). For example, a piece of tape of known length can be placed on the leg of a walking or running persori, so that measurements from the video can be related to those of a known scale (Figure 2). Alternatively if the length of the subject’s limb or, for example, the average wingspan of a bird species is known, then this information may be used to make estimates of measurements taken from the video. Furthermore, placing markers on points of interest (e.g., knee or ankle) will make the points more visible and easier to follow on the video, which will yield more precise measurements. In our walking example, we placed markers on our subject’s hip and knee so we could easily follow the motion of these points and determine the angles formed by the joints (Figure 2). It is important to note that when angles are being measured, the change in the angle must be in the same plane as the camera’s view. If the angle between the camera and the subject is pronounced then the angle and speed measurements will be Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 165 inaccurate. Kinematics^ After collecting the desired footage, begin the analysis by watching a movement in slow motion on a television monitor, or import the video sequences into a computer that has a video capture card and the necessary software to download video. On Screen Analysis One goal of a kinematic study may be to slow a motion so that it can be described or illustrated. An easy way to illustrate the large scale changes that occur during a motion or locomotor cycle is to place an acetate sheet, tracing paper, or some other transparent material on a television monitor and trace sequential images. A video projector can also be used to project and enlarge an image (this is similar to using a film projector; see Carpenter and Duvall 1983). For example. Figures 3 and 4 are traces of a human leg during walking and a dove flying away from a camera, respectively. To illustrate the path of a body part with respect to the body (Figure 5), trace the body’s initial position along with a reference point (e.g., the eye). Then, after advancing one frame, align the reference point(s) on the image to the reference point marked on the acetate sheet, and plot the new positions of the body part of interest (e.g., the wingtip). Figure 5 illustrates the path of the wingtip with respect to the body for a black-billed magpie and a collared dove during slow, level flight; although both of the birds are the same mass and were performing the same task, their kinematics are quite different. To plot the motion of a body part through space, trace the initial body position of the subject and the points of interest, then, without moving the acetate, plot the sequential position of the points of interest for each frame of the motion. For instance, in Figure 6 we plotted the paths of the wrist, the wingtip, and the eye for the same wingbeats illustrated in Figure 5. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 166 Motions also can be quantified by measuring the movement of traced points or angles of interest. In the example of a human walking, it may be of interest to know the velocity or speed of the hip, knee, and ankle during a normal walking cycle. To collect this information plot the position of each point (area of interest) in successive frames and measure the distance between each point (the distance that the point moved). If the distances measured on the tracing can be related to a known scale (e.g., a 30 cm strip of tape on the subject’s leg), then the distance that the points moved in a given period of time can be calculated. For example, if the 30 cm tape on the subject’s leg measures 2 cm on the video monitor, then every centimeter measured on the tracing will be equal to 15 cm on the subject (Note: this is true as long as the subject remained the same distance from the camera and the camera was not zoomed in or out during the recording). Thus, if the subject’s knee moved 4 cm on the video monitor, it actually moved 60 cm. If 4 frames of standard video (30 frames/s) elapsed for the knee to move 60 cm, then the knee moved 60 cm in 4/30 s. Thus, the knee moved at a speed of 450 cm/s or 4.5 m/s, that is: linear speed = distance moved = 60 cm = 450 cm/s = 4.5 m/s time elapsed C4 frames’) (30 frames/s) Further, the acceleration of the knee or another point, can be calculated by measuring the change in the speed of the point divided by the elapsed time^. For example, if the speed of the knee during a forward motion is 4.5 m/s, and the speed of the leg 20 frames later in the same motion is 1.5 m/s, then the average acceleration for the knee is the difference between the initial speed measurement and the final speed measurement divided by the time elapsed. Thus, the knee accelerated at -4.5 m/s, that is: acceleration = Final speed - initial speed = 1.5 m/s - 4.5 m/s = -4.5 m/s- elapsed time (20 frames! (30 frames/s) Moreover, if angle changes for a subject are of interest, then these can be measured Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 167 easily with a protractor. To measure joint angle changes, draw straight lines with a ruler between the points of interest (e.g., from the hip to the knee and from the knee to the ankle) and measure the angle made between the lines with a protractor. Next, advance the video and repeat the measurements at the same joint throughout the selected sequence. These angle changes can then be plotted against time, or they may be used to calculate angular velocity (the change in angle over time) or angular acceleration (the change in angular velocity over time). For example, if the joint angle of the knee changes from 180° to 90° in 0.5 s then the angular velocity can be calculated in the following manner"^: angular velocity = change in angle - 90° - 180°/s = n radians/s elapsed time 0.5 s If the angular velocity measured changed from 180°/s to 250°/s in 2 s then the angular acceleration can be calculated in the following manner: angular acceleration = (final angular velocity - initial angular velocitv ) elapsed time = (250°/s - 180°/s) = 35°/s2 2 s Analysis with a computer Analyses similar to those described above can be done with the aid of a computer with a video capture card and associated software. When capturing video be aware of the following potential problems: • The sampling rate of the capture card may be slower or faster than the rate at which the video plays, which may either cause frames to be unrecorded or to be duplicated, respectively. If the video capture card is sampling slower than the video is being played, try playing the video in slow motion as it is being captured and delete sections of video that are duplicated. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 168 • If memory space on a computer is limiting, rather than save an entire video-clip, only save video-still images that are important for analysis and note the number of frames between each video-still. Once a video section has been captured, open the video clip or selected frames in an image analyzer, such as NIH Image, or Videopoint (Lenox Softworks, © 1997 Mark Luetzelschwab and Priscilla Laws). NIH Image is provided at no charge from the National Institute of Health, and can be downloaded from the worldwide web along with additional image analysis files {http://rsb.info.nih.gov/nih-image/Default.html.). Most versions of NIH Image are produced for Macintosh computers, but a PC version of the program is available through Scion Corp.(http://www.scioncorp.com). To access help using NIH Image or detailed descriptions of the program’s functions refer to the online manual at http://rsb.info.nih.gov/nih-image/manual/contents.html. Videopoint can be ordered for Macintosh or personal computers at http://www.lsw.com/videopoint or http://www.pasco.com. The Videopoint home page also includes sample videos, examples of kinematic projects, classroom activities, example homework assignments and other information on using the program. For additional information on using kinematics and kinematics programs (including several published papers that can be viewed online [Beichner 1994, 1996]) visit Dr. Robert Beichner’s web page at {http://www2. ncsu. edu/ncsu/pams/physics/Physics_Ed/ Authors/Beichner.html ). There are several advantages in using an image analysis program to perform kinematics, such as: • If there is a known scale on the subject or in the video-still, then the program can convert the on screen measurements to life size scale. • An XY coordinate plane is automatically placed on a picture opened in these Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 169 programs. (Note: In order to accurately measure movements in an XY plane the camera should be stationary and the subject should be moving across the camera’s field of view.) • Information on the location of several points and/or the size of several angles can be collected at one time. After collecting the XY coordinates of a moving point, the distance a point moved can be calculated with the Cartesian geometry distance formula. distance = V((X ,-X )^-(Y ,-Y )^) where X = initial X coordinate n+l n ' ' n+1 n You may then use the formulas given in the previous section (On Screen Analysis) to calculate linear speed and acceleration of a point, as well as angular velocity and angular acceleration. If your confidence in the accuracy of the measurements is low, you may want to calculate speeds averaged over many frames and accelerations averaged over many more frames. E xtensions The techniques outlined in this paper can be applied to a variety of topics. Kinematics have been used in University of Montana courses for inquiry based learning, in which students have done such diverse projects as investigating hip dysplasia in dogs to pouncing in cats and pecking in woodpeckers. The number of projects available for inquiry or discussion is limited only by the imagination and, to a lesser extent, by the resources available. Kinematic studies are appropriate at many academic levels, depending on the material being emphasized. Illustrating the motions of animals through tracings can be used in pre-high school classes. Analyses involving velocities, accelerations, and angle changes are appropriate for both physics and biology classes at the high school and college level. Furthermore, kinematic studies lend themselves to poster presentations that model Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 170 communication at scientific meetings. In addition, advanced students may be introduced to complex concepts, such as the effects of body mass on locomotion, or the relationship between gait transitions and posture in animal design (e.g., Biewener 1989). The evolution of stance, and locomotor strategies with the transition from aquatic to a terrestrial environment, and the corresponding morphology all can be investigated further through inquiry or through demonstration. The role of ontogeny (development) on locomotion is also a fruitful area for students to explore; for example, how do kinematic strategies change as a human child (or another young animal) learns to walk? In addition, the role of morphogenesis, such as the development from a tadpole to an adult frog, may be an interesting area in which students can investigate the development and the associated locomotor or feeding functions of various bodyparts in animals. Some examples from the literature include; • Walking and stability in turtles and other tetrapods (Gray 1968, Alexander 1992) • Pendular versus speed walking in humans (Alexander 1992) • Gait changes in mammalian locomotion (Biewener 1989, Alexander and Jayes 1983) • Locomotion in snakes (Gray 1968) • Swimming in penguins (Clark and Bemis 1979). Once an instructor is familiar with the techniques of kinematics, it is possible to lead students through guided and open inquiries on locomotion and animal design. Whether kinematics is used for demonstrations, or inquiry-based laboratories, we have found that bringing kinematics into the classroom leads to considerable discussion and discovery. Acknowledgements Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 171 We thank Dr. Carol Brewer for her comments and recommendations throughout the conception and creation of this manuscript, and Dr. Douglas Warrick for his helpful comments on the manuscript. R eferences Alexander, R.M and A.S. Jayes. 1983. A dynamic similarity hypothesis for the gaits of quadrupedal mammals. Journal of Zoology (London) 201:135- 152. Alexander, R.M. 1992. Exploring Biomechanics. Scientific American Libraries. New York, New York. Balcolmbe, J. 1997. Student/teacher conflict regarding animal dissection. The American Biology Teacher 59(1): 22-25. Beichner, R. 1994. Testing student interpretation of kinematics graphs. American Journal of Physics. 62: 750-762. Beichner, R. 1996. Impact of video motion analysis on kinematics graph interpretation skills. American Journal of Physics. 64: 1272-1278. Biewener, A. A. 1989. Mammalian terrestrial locomotion and size. Bioscience 39(ll):776-783. Biological Sciences and Curriculum Study (BSCS). 1993. Developing Biological Literacy. BSCS Press, Colorado Springs, CO. Carpenter, G.C. and D. Duvall. 1983. Motion Picture and videotape analysis of behavior. The American Biology Teacher 45(6): 349-352. Clark, B.D. and W. Bemis. 1979. Kinematics of swimming of penguins at the Detroit Zoo. Journal of Zoology 188: 411-428. Gray, J. 1968. Animal Locomotion. W.W. Norton and Company, Inc., New York. National Research Council (NRC). 1997. Science Teaching Reconsidered. National Academy Press, Washington D C. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 172 Endnotes 1. If the video has not been collected in the manner described in the text, the velocity or acceleration of an animal or body part with respect to a stationary object can still be estimated, but the calculations and assumptions necessary are beyond the scope of this paper. 2. To demonstrate the process of motion analysis we will present several examples; the examples with birds were collected on Hi-8 video and transferred to SVHS video, while those involving human walking were collected on 8 mm video and transferred to standard VHS videotape (Note: Transferring video from one tape of the same format (S-VHS and HI-8 being similar in format and VHS and 8 mm being similar in format) to another will result in a slight loss of resolution, but in most cases the loss is negligible. However, transfering from S-VHS or Hi-8 format to standard VHS or 8mm will result in considerable loss of resolution.) The highest shutter speeds that produced correctly exposed images were used in all cases; in the bird examples, we used shutterspeeds of 1/1000 s or 1/2000 s and in the human (since the motions during walking are not terribly fast) we used a shutterspeed of 1/500 s. All of the examples were then played on a VCR capable of sampling at 60 frames/s. 3. Due to the nature of calculating acceleration, slight errors in initial measurements can lead to huge measurement errors for acceleration; one way to reduce the error is to average the acceleration over many frames. 4. You may wish to convert measurements in degrees to radians because most spreadsheets use radians in trigonometric calculations. You can use the formula 180° = 7t radians, or 1 radian = 180 °/ti to convert between radians and degrees. A good approximation of radians can also be obtained by dividing the measurement in degrees by 57.2958. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 173 APPENDIX G FIGURE LEGENDS Figure Gl. A hypothetical example illustrating the connection between sampling rate and the information gathered with 1) a slow sampling rate (e.g., 30 frames/s video), 2) a sampling rate twice as high (e.g., 60 frames/s video), and 3) a fast sampling rate (e.g., high speed video or film). Figure G2. Placing a grid behind the subject and/or an object of known scale (strip of white tape) in the field of view or on the subject allow you to convert the measurements taken off of the video to the correct units. In addition, placing markers at points of interest (the hip and knee here) can lead to more accurate measurements. Figure G3. Five tracing of a human walking. The motion was traced once every 15 frames from 60 frames/s video. Figure G4. Sequential tracings of a collared dove during a wingbeat cycle (60 frames/s video) Figure G5. The path of the wingtips with respect to the body in a Ringed Turtle-Dove (left) and a Black-billed Magpie (right) during one wingbeat (60 frames/s video) Figure G6. Tracings of the wingtip, wrist and eye of a Ringed Turtle-Dove (left) and a Black-billed Magpie (right) during a wingbeat cycle (60 frames/s video) Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Figure Gl 174 Figure 0 2 m Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Figure G3 175 Figure G4 9 8 ■s? ' 5 4 3 2 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 176 Figure G5 Figure G6 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 177 APPENDIX H Level flight wingbeat frequencies of Herons (Ardiedae) Considerable discrepancies exist between predicted flight behavior and empirical evidence. Pennycuick (1989, 1990) suggests that during level flight birds beat their wings at a natural frequency (cruising frequency) which requires the least amount of power input. This cruising frequency should be proportional to M' which has been documented among a large group of seabirds (Pennycuick 1990, 1996). During level flight wingbeat frequency scales as M"* in woodpeckers (Tobalske 1996) and across many bird species is scales as (Greenwalt 1962, Van Den Berg and Rayner 1995), yet within the doves it scales as The purpose of this study was to examine the scaling of level flight wingbeat frequency in Herons, a group which has been included in Pennycuick’s (1990, 1996) data sets. Flights used for analysis were recorded in the Rio Grande Valley of southern Texas (elevation <50m) Kinematics were collected using a high speed video camera (60 fields s '\ Hi-8 format, Sony Model 910). The Hi-8 video was transferred to S-VHS and a time code (Horita II model TG 50) was added. Video was viewed and analyzed using a Panasonic AGI 960 editing video player. Flights of herons during level, non-maneuvering, uninterrupted bouts of flapping flight in calm air (wind < 1ms ') were included in the analysis of wingbeat frequency, which was determined by dividing the number of Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 178 wingbeats recorded in a flight by the elapsed time (determined using the number of frames elapsed). To reduce the risk of pseudo replication from measuring a single bird more than once, different geographic locations (areas separated by more than 3 kilometers) were used as sample units for field data, unless several birds were recorded simultaneously, in which case, each bird was considered a sampling unit. The results of this study are summarized in Figure HI. Pennycuick, C.J. 1990. Predicting wingbeat frequency and wavelength in birds. J. Exp. Biol. 150: 171-185. Pennycuick, C.J. 1996. Wingbeat frequency of birds in steady cruising flight: new data and improved predictions. J. Exp. Biol. 199(7): 1613-1618. Scholey, K.D. 1983. Developments invertebrate flight. Ph.D. thesis, Univ. Bristol, United Kingdom. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 179 APPENDIX H FIGURE LEGENDS Figure HI. Scaling of level flight wingbeat frequency in Herons (++ indicates value taken from Pennycuick [1990]; ** indicates value taken from Scholey [1983]). Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 180 Figure HI. 0.7 ■5.0 N K o N g 0.6. .4.0 3 CT U Little blue Heron + + c g Black-crowned Night Heron OJ 3 c3 Yellow-crowned Night Heron cr E I c 0.5 . .3.2 Great Blue Heron ++ y .2.5 .SP Grey Heron (Ardea cinerea) ** 13 > y = -0.260x + 1.265 r"^ = 0.863 E RMA = -0.279 0.3 2.0 2.0 2.5 3.0 3.5 ( 100) (316) ( 1000) (3162) Log Body Mass (g) (Body Mass [g]) Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.